Populism, Illiberalism, and Popular Sovereignty in Latin America
by Julio F. Carrión
There is a contested relationship between populism and democracy. Defenders of populism argue that populism enhances the democratic dimension of liberal democracy (popular sovereignty), even if its liberal component (checks and balances) may suffer as a result. The present paper rejects this claim on two counts: conceptual and empirical. The paper shows that the liberal and democratic dimensions of democracy are deeply interwoven in practice. Effective checks and balances (a central component of the liberal dimension) are necessary for the full exercise of popular sovereignty (the preeminent component of the democratic dimension). This paper shows that populism in power moderates the relationship between checks and balances (measured as horizontal accountability) and popular sovereignty (measured as free and fair elections). Therefore, the paper concludes that when checks and balances are eroded by populist chief executives, so too is the exercise of popular sovereignty. Empirically and conceptually, the liberal dimension of liberal democracy cannot be diminished significantly to enhance the democratic component, as theorists of populism claim, because the weakening of the first leads to the erosion of the second. The modeling strategy is based on a fixed-effect panel design of 19 Latin American countries in the period 1979-2021.
Keywords: populism; popular sovereignty; horizontal accountability; elections; Latin America.
Does populism, when in power, strengthen or erode democracy? And if populism enhances democracy, does it do so at the expense of liberalism? While recent scholarship argues that populism in power always and inexorably erodes democracy, several political theorists hold a divergent view. They have long portrayed populism as a healthy reaction to the elitist tendencies built into liberal democracies—that is, as a force that seeks to recover and strengthen the principle of popular sovereignty, which has, in their view, been diluted by constitutional constraints. Margaret Canovan, for instance, writes that populism is “an appeal to ‘the people’ against both the established structure of power and the dominant ideas and values of the society.” Populist chief executives frequently portray representative institutions as obstacles to popular sovereignty and complain that entrenched elites in legislative, judicial, and oversight bodies represent an obstacle to the full exercise of popular will. Viktor Orbán infamously argued that “a democracy does not necessarily have to be liberal.”
To be fair, defenders of populism point to real challenges that liberal democracies have been unable to address, such as growing income inequality, state capture, and the outsized political influence of economic elites. Ernesto Laclau and Chantal Mouffe have a valid point when they criticize “the system of power” that “redefines and limits the operation” of the values of liberty and equality. But the question is whether reducing liberal protections from enlarged executive power is the best way to combat that “system of power” while still protecting democracy. Nadia Urbinati replies to Laclau and Mouffe by asking an important question: “Do populist constitutions of democracy … include things like civil liberties and the separation of powers?” and, if defeated elites retain the right to reorganize, compete, and perhaps even win elections, then how is populism “any different from Schumpeterian democracy?”
Another strand of scholarship argues that populism and democracy are incompatible, pointing to populism’s anti-pluralism and illiberalism as drivers of democratic regression. However, most of these studies do not empirically analyze the influence of populism in power on the relationship between the liberal and democratic dimensions of existing political regimes. Is populism in power in Latin America able to enhance democracy by reducing the liberal elements of the political regime? Does the erosion of liberal procedures—a process embraced by populism’s defenders—lead to an increase in popular sovereignty? This paper is an effort to answer these questions by exploring the empirical relationship between liberalism and democracy under populist and non-populist chief executives.
In what follows, the liberal dimension of democracy is operationalized as the effective exercise of check and balances (also known as “horizontal accountability”). The democratic dimension of democracy is popular sovereignty, operationalized here as holding free and fair elections. There is general agreement that institutional constraints on the power of the state lead to clean elections. Horizontal accountability, defined as the ability of state agencies to check the behavior of powerholders, is part of an institutional package that produces good-quality democracy. Merkel and Croissant point out that “democratic elections need the support of complementary partial regimes, such as the rule of law, horizontal accountability, and an open public sphere.”Can populism in power, as some argue, strengthen people’s right to free and fair elections while eroding these liberal “complementary partial regimes”?
Proponents of populism argue that liberal procedures can curtail the exercise of popular sovereignty. The alternative view, which I seek to test here, contends that these two dimensions cannot be disentangled without affecting people’s right to free and fair elections. I argue that due to its strong illiberal tendencies, populism in power undermines the exercise of free and fair elections by moderating the impact that horizontal accountability (a liberal dimension) has on the conduct of elections (a democratic dimension). To test this posited moderating impact, I utilize a panel design using Varieties of Democracy data (V-Dem version 12) for 19 Latin American countries. I start the analysis with 1979, when Ecuador held a presidential run-off that inaugurated what came to be known as the third wave of democratization in the region. The data end in 2021.
Before proceeding to the analysis, it is important to offer a brief definition of populism. In the field of comparative politics, two treatments are frequently used. The first is the “ideational approach,” which sees populism as a “thin ideology” that divides the political world into two: a noble and pure “people” confront a morally corrupt “elite” that denies the enactment of the general will. The second defines populism as a political strategy to seek and exercise power. This strategy relies on establishing unmediated, direct links between a personalistic leader and a largely unorganized mass of followers. I embrace the political-strategic approach to populism because it captures the main domain of populism, which is power and domination, as Weyland puts it. At the same time, it is indisputable that populist leaders see the world in a Manichean fashion, embracing an “us-versus-them” mentality that leads them to reject pluralism and downplay the importance of institutional constraints on the exercise of power.
I therefore define populism as a political strategy that seeks to establish direct, unmediated links with generally unorganized followers, embraces an “us-versus-them” mentality, and exhibits deep distrust of checks and balances. In the paper, populism in power is measured by a hand-coded dummy variable that assigns a 1 for each year (or partial year) a populist was in power, and 0 otherwise. (This is discussed further in the section on populism in Latin America below.)
The paper proceeds as follows. The first section revisits the debate regarding the relationship between populism and the liberal and democratic components of liberal democracy. It argues that the illiberal tendencies associated with populism endanger liberal practices and, in the process, the democratic dimension of existing democracies. The second section briefly describes the experience of populism in power in Latin America, showing the diversity of outcomes for democratic rule the states of the region exhibit. The third section hypothesizes the empirical relationship between populism, horizontal accountability, and democratic elections. The fourth section discusses the results. The concluding section offers some final thoughts about the relationship between populism, illiberalism, and popular sovereignty.
Populism, Liberalism, and Popular Sovereignty
Theorists of democracy have long identified two principles, or dimensions, that underpin contemporary democracies. The first is popular sovereignty, manifested in the ideal of majority rule. This principle embodies the notion that people are the source of political legitimacy and the foundation of the political regime. The second refers to the liberal rights—civil and political—that every citizen enjoys as a member of the political community. The liberal component is manifested in the practice of constitutionalism, which not only guarantees individual rights, but also limits the power of the state. Some argue that the coexistence of liberal and democratic principles in contemporary liberal democracies is both a historical accident and an unexpected combination produced by the demands of practical politics. This paradoxical confluence was, according to Mouffe, the product of bitter historical battles. This conventional view, which sees the liberal and democratic principles cohabitating in tension, is described as the “two-strand model” of liberal democracy.
Theorists advocating for populism operate within this two-strand conceptual universe of liberal democracy. Nadia Urbinati is correct in pointing out that those who praise populism do so precisely because it is meant to reinforce the democratic pillar of liberal politics that has been eroded in contemporary democracies. Indeed, some argue that populism reclaims, revives, and strengthens the popular sovereignty dimension that is very limited in today’s democracies. Margaret Canovan, while clear-eyed about its risks, suggests that populism can realize the promises of popular sovereignty that liberal democracy offers but on which it all too often fails to deliver. Populism, then, recurs because it resides in the gap between the “two faces” of democracy: its “redemptive face” (the promise of liberation through popular power) and its “pragmatic face” (a form of government that manages conflict). Yves Mény and Yves Surel argue that populism has “a proximity to, and affinity with, the democratic principle.” Ernesto Laclau, referring to their work, says that “populism is the democratic element in contemporary representative systems” (emphasis added).
A related perspective proposes that populism can be both a corrective and a threat to democracy, for it strengthens participation and inclusiveness (popular sovereignty) while weakening public contestation (a liberal component). In this version, the impact of populism is disaggregated in relation to liberal democracy’s two dimensions: populism is beneficial to the democratic face of democracy but pernicious to its liberal component. In a similar vein, Mudde asserts that populism is not hostile to representative democracy as such, but rather to the institutions of liberal democracy.
This view that popular sovereignty is strengthened by populism, even if this comes at the expense of the liberal component, has been challenged. In an important contribution, and drawing on the work of Rawls, Habermas, and Lefort, Rummens argues that the “two strands” or “two pillars” understanding of liberal democracy is flawed because it assumes that one can function properly without the other. For Rummens, liberal democracy is not a paradoxical regime that contains two pillars in tension; on the contrary, and relying on Habermas’ language, there is a “co-originality” between them. By this he means that “the liberal and democratic dimensions are not incompatible at all, but represent, rather, inseparable or ‘co-original’ aspects of a regime which aims to preserve and protect human freedom.” Rummens admits he is offering not a historical account but a conceptual argument about the indissolubility of these two dimensions of liberal democracy.
I argue that the threat that populism poses to democracy originates not from its exultation of popular sovereignty and majority rule but from its illiberalism. Populism’s illiberalism is grounded in two claims. The first is the monopoly of representation. This claim emanates from the way populism defines “the people,” which is different from the way liberalism defines this notion. In a liberal democracy, “the people” is understood as “an irreducible plurality, consisting of free and equal citizens;” it is a “unity-in-diversity.” Populism, by contrast, sees “the people” as an ideal homogeneous community that possesses a single general will. Populist leaders claim to be the only ones who represent the people; to that end, they embrace “an exclusionary form of identity politics …[that] tends to pose a danger to democracy.” This claim to a monopoly on representation is the result of a dual move: the artificial creation of the people and the subsequent delegitimization of those that it excludes. Indeed, Arato and Cohen argue that the logic of populism is to depict one part of the people as representative of the whole and to conceive of this imaginary “people” as an ideal unity confronting a dangerous enemy. Moreover, populism rejects “otherness” and therefore delegitimizes any challenge that disrupts the unity of that idealized people.
This delegitimization is manifested in the “us-versus-them” mentality that populism embraces. Given their claim to be the only ones who truly represent the real people, populist leaders can never accept, as Panizza tells us, that “the people’s will can never be fully enacted” and therefore “there could only be contested versions of who the people are, and who has the right to speak on its behalf.” Along similar lines, Abts and Rummens emphasize that the will of the people has to be mediated and is an ongoing construction. In short, the anti-pluralism that many scholars identify in populism is the consequence of its claim to be the sole representative of an idealized people. Stressing that the liberal cannot be disentangled from the democratic, Panizza writes that the “argument for the toleration of differences is not only a liberal argument but a democratic argument as well.” Populism’s illiberalism undermines its claims to be democratic.
The second claim that makes populism illiberal is its view that “constituent power” has unlimited primacy over “constituted power.” Many have noted the predilection of populist leaders for embracing “foundational” or “refoundational” language to imply that they are (re)creating a new political order to recover popular sovereignty undermined by previous governments. It is true that left-wing populist leaders choose to enact their agendas through constitution-making rather than revolt or violence. In fact, it has been argued that there is a populist constitutionalism that relies on the primacy of constituent power to not only enact constitutions, but also reform them at will, thus situating politics higher than law. Others, more critical of the motives behind wholesale constitutional changes, see them as part of an effort by chief executives to “fix” the political game and codify power asymmetries that favor them.Regardless of whether there is such a thing as “populist constitutionalism,” populism rejects the principle that constitutions are written not only to organize power (with the people as the source of their legitimacy), but also to set limits on those who wield power in the name of the people. Many courts, captured by populist leaders and working under the constitutions these leaders have enacted, operate not to limit executive power but to protect and enhance it. The concepts of “autocratic legalism,” “discriminatory legalism,” and “abusive judicial review” illustrate such instrumental use of constitutionalism under populism. This is also manifested in the frequency with which populist leaders amend their own constitutions. Once they have acquired enough power asymmetry, they engage in practices that would have been unthinkable when they first came to power. Jan-Werner Müller concludes that populists might draft constitutions, but “they violate certain core ideas of a normative understanding of constitutionalism” (emphasis original). This tendency to see constitutionalism as an instrument for enhancing—rather than limiting—the power of those in charge of the state underlies the illiberal practice of populism in power.
Building on the previous discussion, I argue that populism in power erodes the democratic (popular sovereignty) dimension of democratic rule precisely because it reduces the effectiveness of liberal practices that guarantee its exercise. I contend that populism moderates (decreases) the positive impact of horizontal accountability on free and fair elections. Before proceeding to the empirical section, a short discussion of the prevalence of populism in Latin America is in order.
Populism in Latin America
Populist movements and leaders claim to represent people’s will when they confront the strictures posed by liberal institutions (for example, the need to have legislative supermajorities and overcome judicial review to enact significant constitutional reform). Yet in Latin America, as elsewhere, populism’s record of protecting people’s right to participate in free and fair elections is mixed. While it is true that populism in power does not always lead to a significant decline in levels of democracy, there are more than a few cases in Latin America where it has done so. Case studies show that in such places as Venezuela, Peru, Ecuador, Bolivia, and Nicaragua, populist leaders tilted the electoral playing field to such an extent that elections held under their rule could not be called free or fair. By contrast, in other cases—among them Argentina, Brazil, and Panama—the arrival of populism in power did not produce a noticeable change in the quality of elections. In a similar vein, while some populist leaders are successful in establishing political monopolies, many others fail to do so, facing early political deaths or eventual collapse.
Of the subregions in Latin America, the Andes are perhaps the one that has been most impacted by populist governments since 1979. During the period under study, every single country in this subregion elected at least one populist presidential candidate. Some did not last long, as opposition-controlled legislatures impeached and removed them, but others were able to change their countries’ constitutions to secure extended reelections. Lucio Gutiérrez was elected in Ecuador in 2003 on a platform of economic reform and clean government. He had the support of the indigenous movements and left-wing parties. After clashing with the opposition-controlled legislature, he was removed from office in 2005. His government had no opportunity to influence the nature of the electoral game. Rafael Correa, elected in late 2006, remained in office for a decade. In the process, he built a regime that many labeled competitive authoritarian. In Peru, Alberto Fujimori shut down Congress and the judiciary in 1992; in 2000, he held elections that were considered neither free nor fair. Evo Morales stayed in power in Bolivia from 2006 until 2019, when he was removed by a popular uprising protesting a rigged electoral process. In Venezuela, Hugo Chávez assumed office in 1999; his hand-picked successor, Nicolás Maduro, is still in power. This regime is widely considered to be full authoritarian. In Colombia, Álvaro Uribe was elected in 2002. He served two four-year terms and tried to change the constitution to run for a third term but was prevented from doing so by the Constitutional Court.
Most of the Southern Cone countries have less experience with populism in power than the rest of the region, although Argentina and Brazil are notable exceptions. Argentina is one of the birthplaces of Latin American populism: Juan Domingo Perón was first elected in 1946. Since its return to democracy in 1983, Argentina has more often than not been ruled by populist presidents: Carlos Menem (1989-1999); Nestor Kirchner (2003-2007); and his widow, Cristina Fernández de Kirchner (2007-2015). In Brazil, the first democratic elections conducted after the end of the military government were held in 1989. They were won by the young, charismatic populist Fernando Collor de Mello. He soon confronted allegations of corruption, with the result that the Senate started impeachment proceedings. Populism returned in 2018 with the election of the right-wing populist Jair Bolsonaro. He failed to secure reelection in 2022. In Paraguay, the short-lived presidency of Fernando Lugo (2008-2012) marks the country’s only experience of a populist in power. Neither Chile nor Uruguay elected a populist president in the period under study. Notably, no single populist chief executive in this subregion undermined in significant ways the democratic character of their national elections.
With the exception of Costa Rica all Central American countries experienced populism between 1979 and 2021. Guatemala elected the populist Jorge Serrano Elías in 1991. Once in office, Serrano tried to imitate Peru’s Fujimori by declaring an auto-coup. A swift response from civil society and other political actors defeated this power grab and forced Serrano to leave the country. Guatemala had a second encounter with populism when Álvaro Colom was elected in 2008 on a platform to reduce poverty. In Honduras, populist Manuel Zelaya assumed office in 2006. In 2009, he pushed for a referendum to change the constitution, which led to a confrontation with the other branches of government. When he refused to follow a judicial ruling, the Supreme Court issued an order for his arrest and the Army removed him from power. In Honduras, subversion of the constitution to enable an illegal reelection occurred under Juan Orlando Hernández, a non-populist and leader of the conservative National Party.
In Nicaragua, the election of the populist Daniel Ortega in 2006 opened a period of growing autocratization akin to the one observed in Venezuela. Ortega’s total control of institutions, including the judiciary, has led to the emergence of full-scale authoritarianism in the country. In Panama, by contrast, the presidency of the populist businessman Ricardo Martinelli did not affect the level of democratic elections, even though Martinelli was involved in a series of corrupt practices, for which he was indicted after he left office in 2014. In El Salvador, populism came to power rather recently: the country elected Nayid Bukele in 2019. Recent events there do not bode well for the survival of popular sovereignty in that country. Elections are not scheduled until 2024, but the country’s Supreme Court has already enabled Bukele to run for immediate reelection. In Costa Rica, the populist Rodrigo Chaves assumed office in 2022, too recently to be included in the V-Dem data set. Finally, in Mexico, left-wing populist Andrés Manuel López Obrador won the 2018 election after twice running unsuccessfully. His administration has faced widespread criticism due to his harsh anti-press rhetoric and his efforts to influence the judiciary. He has announced that he does not intend to seek reelection—a possibility formally precluded by the Mexican Constitution.
In the statistical analysis that follows, I hand-code the variable “populism in power” as a dummy to account for the presence or absence of a populist chief executive. The decision to classify a chief executive as populist is based on an analysis of the literature (see source in Table 1) and complemented by the author’s judgment. Every year (or partial year) that a populist was in power is assigned a 1; otherwise, the year is coded as 0. Table 1 reports the list of populist presidents from 1979 to 2021, showing how frequently Latin American voters have selected a populist candidate as chief executive. A cursory look at the names suggests that only in a handful of cases has democracy come to an end as a result of the populist experiment. However, even in those cases where democracy has survived, the country’s institutionality has been affected, as some of these populist presidents have been removed by impeachment or coups.
Table 1. List of Latin American Populist Presidents, 1979-2021
|Country||Populist Presidents (Years)|
|Argentina||Carlos Menem (1989-1999); Nestor Kirchner (2003-2007); Cristina Fernández de Kirchner (2007-2015)|
|Bolivia||Evo Morales (2006-2019)|
|Brazil||Fernando Collor de Mello (1990-1992); Jair Bolsonaro (2019-2021)|
|Colombia||Álvaro Uribe (2002-2010)|
|Dominican Republic||Hipólito Mejía (2000-2004)|
|Ecuador||Abdalá Bucaram (1996-1997); Lucio Gutiérrez (2003-2005); Rafael Correa (2007-2017)|
|El Salvador||Nayib Bukele (2019-2021)|
|Guatemala||Jorge Serrano Elías (1991-1993); Álvaro Colom (2008-2012)|
|Honduras||Manuel Zelaya (2006-2009)|
|Mexico||Andrés Manuel López Obrador (2018-2021)|
|Nicaragua||Daniel Ortega (2006-2021)|
|Panama||Ricardo Martinelli (2009-2014)|
|Paraguay||Fernando Lugo (2008-2012)|
|Peru||Alberto Fujimori (1990-2000)|
|Venezuela||Hugo Chávez (1999-2013); Nicolás Maduro (2013-2021)|
Source: Compiled by the author on the basis of Houle and Kenny 2016; Huber and Schimpf 2016; Ruth-Lovell, Lührmann, and Grahn 2019; Weyland 2020.
In the next section, I analyze how populism in power affects the relationship between horizontal accountability and free and fair elections.
Populism, Horizontal Accountability, and Free and Fair Elections: Hypotheses
As noted in the introduction, there is general agreement that horizontal accountability is requisite for free and fair elections. Checks and balances are frequently mentioned as necessary (although certainly not sufficient) conditions for reducing the probability of rigged or unfair elections. The V-Dem Institute offers two indices that allow us to examine the empirical relationship between checks and balances and free and fair elections. The Clean Elections Index (CEI) (v2xel_frefair), which I use as an indicator of free and fair elections, is a 0-1 summary index built on country experts’ answers to a set of questions that probe the fairness and cleanliness of elections. As an indicator of checks and balances, I use the V-Dem Horizontal Accountability Index. As described by V-Dem, this 0-1 index “concerns the power of state institutions to oversee the government by demanding information, questioning officials and punishing improper behavior.” These institutions include the judiciary, the legislature, and oversight bodies.
Large quantitative studies of populism in power show that it tends to undermine democracy. Studies that rely on medium-N analyses tend to recognize that not all populism in power leads to the deterioration of democracy or competitive elections. They therefore prefer to focus on which conditions might create opportunities for populist chief executives to subvert democracy. Studies of specific cases of populism in power obviously tend to pay attention to those where democracy is severely eroded. Only rarely do case studies focus on situations where the worst consequences of populism in power on democracy have been averted.
While Table 1 and some of the small-N and medium-N analyses suggest a variety of relationships between populism in power and the conduct of free and fair elections, the literature quoted above tends to argue that populism in power will directly affect existing levels of democracy and, therefore, the democratic character of elections. Given that my main goal is to determine whether populism in power will have a moderating impact on the relationship between horizontal accountability and free and fair elections, my first hypothesis is framed as a null hypothesis on the direct impact of populism in power on elections when controlled by the effects of horizontal accountability.
Hypothesis 1: Populism in power will not have a significant impact on the levels of free and fair elections at average levels of horizontal accountability
Because my interest here is in assessing the influence of populism in power when the functioning of horizontal accountability is at a regional average, I center the horizontal accountability variable at its mean. The question, therefore, is straightforward: What is the individual effect of populism in power on the Clean Elections Index when horizontal accountability has a value equal to its mean?
The argument advanced in this paper is that populism in power tends to moderate the strong impact that effective checks and balances have on the existence of free and fair elections. Therefore, the paper’s second main hypothesis is:
Hypothesis 2: Populism in power will have a moderating impact on the effect that horizontal accountability has on conducting free and fair elections. That is, when populism is in power, checks and balances will be less effective in determining free and fair elections than when populism is not in power
I test this hypothesis using the interaction term between populism in power (a dummy variable) and horizontal accountability (centered at its mean). The expectation is that the slope of the relationship between horizontal accountability and elections will be much steeper when populism is not in power than when it is.
To test these hypotheses, I employ a time-series cross-sectional regression with fixed effects. The total number of observations is 817, of which 153 are coded as populist years (18.8 percent of all observations). I choose a fixed effects model, as a Hausman test for random effects showed that this was the most appropriate strategy. I follow a minimalist modeling strategy to account for control variables. I include as controls only three structural variables: the impact of natural resources extraction (measured as total petroleum, coal, and natural gas production per capita), the average years of education of citizens older than 15, and economic development (measured as GDP per capita). Some of these variables lack data for some years. Except the variable of populism in power (which was hand-coded by the author), all data come from V-Dem dataset Version 12 (2022).
Results and Discussion
Model 1 is the baseline and estimates only the main predictors of interest: populism in power and horizontal accountability (centered at its mean). The three control variables for natural resources extraction, average education, and economic development are also introduced in the baseline model. Again, the dependent variable is the V-Dem Clean Elections Index. This model tests specifically the first hypothesis of this paper. The results are displayed in Table 2. As can be seen, when horizontal accountability is controlled for, populism in power does not have a statistically significant impact on the nature of free and fair elections. This lends support to the paper’s first hypothesis.
The results of Model 1 indicate that when the value of horizontal accountability is equal to its mean, the contribution of populism in power to the CEI is 0.079 and is not statistically significant. This confirms our intuition that the presence of a populist leader in government, by itself, does not have an impact on the cleanliness or fairness of elections. Populism in power does not significantly improve the democratic character of elections, nor does it affect them in a negative fashion. To have a negative impact on elections, populism in power needs to take control of state institutions, which entails reducing the existing checks and balances. Of the control variables, the only one that emerges as a strong influence is the level of economic development, which has the expected sign and is statistically significant.
Table 2. Main Results Fixed Effects Model | Dependent Variable = Clean Elections Index
|VARIABLES||Fixed Effects||Fixed Effects||Fixed Effects||Fixed Effects|
|Populism in power||0.079||0.080||0.056||0.061|
|Horizontal accountability (c)||0.852***||0.936***||0.672***||0.689***|
|Populism in power*horizontal accountability (c)||-0.488***||-0.421**||-0.458***|
|Freedom of association (thick index)||0.313*||0.196|
|Legislative party cohesion||0.081**|
|Natural resources (oil, coal, gas) income per capita||-0.000||-0.000||-0.000||-0.000|
|Average years of education (15+)||0.037||0.018||0.009||0.029|
|GDP per capita||0.025*||0.024*||0.023*||0.011|
|Number of countries||19||19||19||19|
Robust standard errors in parentheses *** p<0.005, ** p<0.01, * p<0.05
Model 2 offers the first test of the second hypothesis. Only the two main effects (populism in power and horizontal accountability) and their interaction term are included in this model alongside the three control variables. The results, once again, emerge as expected. Populism in power, by itself and controlling for the effects of horizontal accountability, does not have a statistically significant influence of the quality of elections. Horizontal accountability, by contrast, is highly statistically significant at the 0.01 level or better. This means that an increase in the values of the horizontal accountability index is associated with a significant increase in the score of the CEI. This is not surprising, since we already know that effective checks and balances are highly correlated with free and fair elections. What is remarkable is that the interaction term between populism in power and horizontal accountability emerges with a negative sign and is statistically significant at the 0.005 level.
This supports H2’s expectation that populism in power will moderate the influence of horizontal accountability on free and fair elections. The negative value of the interaction term suggests that the slopes of horizontal accountability and free elections not only are different but also intersect. This means that populist chief executives will interfere with the institutions in charge of overseeing that free and fair elections are conducted flawlessly. The argument underlying H2 is that if populist chief executives are successful in undermining the liberal mechanisms built into existing democracies, then the exercise of popular sovereignty will suffer as a result. When populism is in power, the ability of horizontal accountability to generate free and fair elections is, on average, curtailed.
Models 3 and 4 test the moderating impact of populism in power found in Model 2 by adding two different sets of predictors that could affect the values of the CEI. Model 3 introduces the V-Dem measure of associational life (v2x_frassoc_thick). The variable measures “[t]o what extent parties, including opposition parties, allowed to form and to participate in elections, and to what extent are civil society organizations able to form and to operate freely?”
The extensive literature on social capital argues that a strong associational life is a predictor of the level of democracy. I therefore include a measure of social capital to see how it affects the relationship between our two main predictors and their interaction term. Model 4 is our full model and includes two additional measures that try to ascertain the possible impact of political polarization and legislative-party cohesion on free and fair elections. In recent years, a growing literature has mentioned political polarization as a possible factor that causes or enables the rise of populist leaders. The argument is that polarization generates strong emotional attachments that exacerbate resentments and make compromise more difficult. This could influence elections because polarization is an incentive to rig elections to favor a given party or candidate that is considered unacceptable by the opposing group. The other variable, legislative-party cohesion, asks: “Is it normal for members of the legislature to vote with other members of their party on important bills?” The assumption is that legislative-party cohesion is an indicator of party health, and therefore of the political system as such, making it a positive influence on the quality of elections. The results of Model 4 in Table 2 are displayed graphically in Figure 1.
Figure 1. Main Results Fixed Effects Models | Dependent Variable = Clean Elections Index
Figure 2 (also based on Model 4, Table 2) shows the substantive effects of horizontal accountability (centered at its mean) on the scores of the Clean Elections Index. The values of the X-axis range from one standard deviation below the mean (-.285) to one standard deviation above the mean (+.285). This figure confirms a strong positive association between the values of horizontal accountability and the scores of the CEI. It clearly suggests that the liberal component (checks and balances) goes hand in hand with the popular sovereignty component (free and fair elections).
Figure 2. Substantive Effects of Horizontal Accountability (centered) on Clean Elections Index
Both Model 3 and Model 4 also confirm the paper’s main hypothesis, namely that populism in power moderates the positive impact of horizontal accountability on free and fair elections. In all models, the interaction term emerges as highly statistically significant and with a negative sign. This means that populism in power has a moderating impact on the effectiveness of existing checks and balances at guaranteeing that democratic elections will be held. The difference between populism and liberal checks on state power when it comes to the exercise of popular sovereignty, as manifested in free elections, is displayed graphically in Figures 3 and 4 (using the results of Model 4, Table 2).
Figure 3 displays the predictive margins (partial effects) of populism in power at one standard deviation below the mean of the centered horizontal accountability variable (-0.285), at the mean value (0), and at one standard deviation above the mean of horizontal accountability (.0285). The slopes intersect because under “no populism in government” the effect of horizontal accountability on the CEI scores is stronger than under “populism in power.” In fact, the slope of the relationship between horizontal accountability and free and fair elections under “populism in power” (solid line) is flatter than the slope under “no populism in power” (dashed line). This suggests that at low levels of horizontal accountability, populist candidates require a higher level of free and fair elections to come to power—higher than a non-populist candidate needs—but once in power, populist governments do not strengthen the quality of elections as significantly as non-populist governments do.
Figure 3. Predictive Margins of Populism in Power on CEI at Representative Values of Horizontal Accountability
Another way to appreciate the moderating effect of populism in power on the relationship between liberal constraints and free and fair elections is shown in Figure 4. This figure displays the average marginal effects of horizontal accountability on the scores of the Clean Elections Index in the absence or presence of populism in power. The information is very clear: switching from “no populism” to “populism in government” reduces the average marginal effect that horizontal accountability exerts on the conduct of free and fair elections (from 0.689 to just 0.230). This means that the positive role liberal institutions play in securing elections as a manifestation of the popular will tends to be undermined when populism is in government. Clearly, when in power, populism on average disrupts the ability of checks and balances to provide a level playing field that would result in free and fair elections.
Figure 4. Average Marginal Effects of Horizontal Accountability on CEI Scores under Conditions of Populism in and out of Power
To test the robustness of the paper’s main findings, I re-estimate Model 4 adding additional controls related to the condition of civil rights in the country and additional measures of social modernization. These additional variables likewise come from V-Dem dataset version 12: the civil liberties index (v2x_civlib), a measure of existing educational inequality for people aged 15 and older (e_peedgini), the average years of education among citizens older than 15 and older (e_peaveduc), and life expectancy at birth (e_pelifeex). The coefficients of the new model, with these additional variables, are displayed graphically in Figure 5. The results confirm the initial findings: first, populism in power does not emerge as a statistically significant predictor of scores on the CEI; second, horizontal accountability is again a strong positive and statistically significant influence on CEI scores; third, as hypothesized in this study, the interaction term for populism in power and the index for horizontal accountability is negative and statistically significant. This confirms that populism in power plays a significant role in ameliorating the impact of checks and balances on the conduct of free and fair elections. Of the other additional variables, legislative-party cohesion emerges as a positive predictor of CEI scores, as it did in Model 4.
Figure 5. Main Results Fixed Effects Models | Dependent Variable = Clean Elections Index (Robustness Test)
Figure 6 shows that the average marginal effects of horizontal accountability remain relatively unchanged in relation to Figure 5. The results indicate that the average marginal effect of horizontal accountability of free and fair elections is 0.675 when there is no populism in power, compared to just 0.206 when populism is in government.
Figure 6. Average Marginal Effects of Horizontal Accountability on CEI Prediction under Conditions of Populism in and out of Power (Robustness Test)
Consistent with the theoretical contributions made by Rummens and by Panizza, this paper shows that there is a clear empirical relationship between the liberal and the democratic (popular sovereignty) dimensions of democracy. I show that a significant practice associated with constitutional liberalism—namely, the presence of checks and balances and horizontal accountability—is a central mechanism to guaranteeing the effective exercise of popular sovereignty, as manifested in the conduct of free and fair elections. Contra those who extol the democratic virtues of populism, the paper finds no evidence that populism in power does more than non-populist governments to increase the democratic exercise of popular sovereignty, at least as measured by the ability of people to select their representatives in free and fair elections.
However, nor does the paper find any evidence that populism in power, by itself and directly, is a cause of the decline of free and fair elections. As some have shown using different approaches and cases, populism in power can affect the effective functioning of democracy and free elections, but that happens under given conditions, and the effect is not universal. This does not mean, however, that the presence of populism in power is without problems. This paper shows that when a populist chief executive is in charge, the positive effects of horizontal accountability on free and fair elections are attenuated. The effect is consistent, statistically significant, and robust.
The general conclusion is that there is no democratic trade-off to be obtained with populism in power: when the liberal constraints on the exercise of power are weakened, so too are the foundations that guarantee the free and fair expression of popular sovereignty. The practice of democracy in Latin America since 1979 shows that electoral democracy (defined as holding free and fair elections) improves when the mechanisms of horizontal accountability are effective. The liberal and the democratic cannot be disentangled in practice: lower levels of horizontal accountability are associated with lower scores on free and fair elections, while higher horizontal accountability is associated with higher values on the V-Dem Clean Elections Index. This study lends support to what such scholars as Urbinati, Arato and Cohen, and Müller, among others, suggest: that illiberal political power’s quest to represent a mythical or “authentic people” can lead down the path to authoritarian government.
Since populism in power moderates the impact of horizontal accountability on free elections, we can also conclude that popular sovereignty suffers in the presence of illiberalism. Populism, as I argue at the outset, is illiberal on two grounds: its claim to a monopoly on representation and its instrumental use of majorities to change constitutional constraints on the power of the executive. This illiberalism does not enhance the democratic dimension; on the contrary, it can generate antidemocratic consequences by leading to the delegitimization of opponents and the undermining of horizontal accountability. Empirically, as well as conceptually, populism cannot enhance democracy at the expense of liberalism. This paper has not examined the impact of populism in power on regime change, nor has it discussed the conditions that could lead to the effective removal of liberal constraints on a populist chief executive. But the findings suggest that when these liberal constraints are compromised, so too are the foundations of free and fair elections.
 Jan-Werner Müller, What Is Populism? (Philadelphia, PA: University of Pennsylvania Press, 2016); William A. Galston, Anti-Pluralism: The Populist Threat to Liberal Democracy (New Haven, CT: Yale University Press, 2018); Takis S. Pappas, Populism and Liberal Democracy: A Comparative and Theoretical Analysis (New York: Oxford University Press, 2019).
 For an excellent analysis of how advocates of populism see it as “an answer to a formal conception of democracy,” see Nadia Urbinati, Me the People. How Populism Transforms Democracy (Cambridge, MA: Harvard University Press, 2019).
 Margaret Canovan, “Trust the People! Populism and the Two Faces of Democracy,” Political Studies 47, no. 1 (1999): 2-16, 3, https://doi.org/10.1111/1467-9248.00184.
 Alberto Fujimori, explaining his confrontational stance toward Congress, said “that there should be no sacred cows in Peru, that no one should have a crown in this country, not even the members of parliaments or the president” (quoted in Charles Kenney, Fujimori’s Coup and the Breakdown of Democracy in Latin America [Notre Dame, IN: University of Notre Dame Press, 2004], 179). Rafael Correa, when first elected in Ecuador, argued in favor of a constituent assembly because “an anti-people majority ha[d] been formed” in the elected Congress (“Ecuador. Correa defiende en Guayaquil su proyecto de asamblea constituyente,” Notimérica, January 4, 2007, https://www.notimerica.com/politica/noticia-ecuador-correa-defiende-guayaquil-proyecto-asamblea-constituyente-20070104052508.html.
 Quoted in Marc F. Plattner, “Illiberal Democracy and the Struggle on the Right,” Journal of Democracy 30, no. 1 (2019): 5-19, 9, https://doi.org/10.1353/jod.2019.0000.
 I thank one anonymous reviewer for pointing out these real problems with the performance of liberal democracies.
 Ernesto Laclau and Chantal Mouffe, Hegemony and Socialist Strategy: Towards a Radical Democratic Politics. Second Edition (London-New York: Verso), xvi.
 Urbinati, Me the People, 35.
 Steven Levitsky and James Loxton, “Populism and Competitive Authoritarianism in the Andes,” Democratization 20, no. 1 (2013): 107-136, https://doi.org/10.1080/13510347.2013.738864.; Robert A. Huber and Christian H. Schimpf, “Friend or Foe? Testing the Influence of Populism on Democratic Quality in Latin America,” Political Studies 64, no. 4 (2016): 872-889, https://doi.org/10.1111/1467-9248.12219.; Saskia Pauline Ruth, “Populism and the Erosion of Horizontal Accountability in Latin America,” Political Studies 66, no. 2 (2018): 356-375, https://doi.org/10.1177/0032321717723511.; Pappas, Populism and Liberal Democracy; Saskia P. Ruth-Lovell, Anna Lührman, and Sandra Grahn, “Democracy and Populism: Testing a Contentious Relationship,” V-Dem Working Paper 2019:91 (2019), https://www.v-dem.net/media/filer_public/a8/b9/a8b9f007-37fd-4f67-8955-f60e11bfef08/working_paper_91.pdf; István Benedek, “Riders on the Storm: The Role of Populism in the Global Crisis of Democracy and in the Functioning of Electoral Autocracies,” Politics in Central Europe 17, no. 2 (2021): 197-225, https://doi.org/10.2478/pce-2021-0009.; but for a more nuanced argument, see Julio F. Carrión, A Dynamic Theory of Populism in Power: The Andes in Comparative Perspective (New York: Oxford University Press, 2022) and Kurt Weyland, “Populism’s Threat to Democracy: Comparative Lessons for the United States,” Perspectives on Politics 18, no. 2 (2020): 389-406, https://doi.org/10.1017/s1537592719003955.
 Guillermo O’Donnell, “Delegative Democracy,” Journal of Democracy 5, no. 1 (1994): 55-69, https://doi.org/10.1353/jod.1994.0010.
 Leonardo Morlino, “What is a ‘Good’ Democracy?” Democratization 11, no. 5: 10-32, https://doi.org/10.1080/13510340412331304589.
 Wolfgang Merkel and Aurel Croissant, “Conclusion: Good and Defective Democracies,” Democratization 11, no. 5 (2004): 199-213, 199, https://doi.org/10.1080/13510340412331304651.
 Cas Mudde and Cristóbal Rovira Kaltwasser, Populism: A Very Short Introduction (New York: Oxford University Press, 2017).
 Kurt Weyland, “Clarifying a Contested Concept: Populism in the Study of Latin American Politics,” Comparative Politics 34, no. 1 (2001): 1-22, https://doi.org/10.2307/422412.
 Ibid., 11.
 David Held, Models of Democracy. Third edition (Stanford, CA: Stanford University Press, 2006).
 Margaret Canovan, “Taking Politics to the People: Populism as the Ideology of Democracy,” in Democracies and the Populist Challenge, ed. Yves Mény and Yves Surel (Basingstoke: Palgrave), 25; Paul Blokker, “Populist Constitutionalism,” in Routledge Handbook of Global Populism, ed. Carlos de la Torre (New York: Routledge, 2019), 116.
 Yves Mény and Yves Surel, “The Constitutive Ambiguity of Populism,” in Democracies and the Populist Challenge, ed. Yves Mény and Yves Surel (London: Palgrave, 2002), 7.
 Chantal Mouffe, The Democratic Paradox (London: Verso, 2000).
 Koen Abts and Stefan Rummens, “Populism versus Democracy,” Political Studies 55, no. 2 (2007): 405-424, https://doi.org/10.1111/j.1467-9248.2007.00657.x.
 Urbinati, Me the People, 571.
 Canovan, “Trust the People!”
 Mény and Surel, “The Constitutive Ambiguity of Populism,” 6.
 Ernesto Laclau, On Populist Reason (London, New York: Verso, 2005), 176.
 Cristóbal Rovira Kaltwasser, “The Ambivalence of Populism: Threat and Corrective for Democracy,” Democratization 19, no. 2 (2012): 184-208, 200, https://doi.org/10.1080/13510347.2011.572619.
 Cas Mudde, “The Populist Zeitgeist,” Government and Opposition 39, no. 4 (2004): 541-563, 561, https://doi.org/10.1111/j.1477-7053.2004.00135.x.
 Stefan Rummens, “Populism as a Threat to Liberal Democracy,” in The Oxford Handbook of Populism, ed. Cristóbal Rovira Kaltwasser, Paul Taggart, Paulina Ochoa Espejo, and Pierre Ostiguy (New York: Oxford University Press, 2017).
 Ibid., 556.
 Ibid., 554.
 Ibid., 558.
 Andrew Arato and Jean L. Cohen, Populism and Civil Society: The Challenge to Constitutional Democracy (New York: Oxford University Press, 2021); Galston, Anti-Pluralism; Mudde and Kaltwasser, Populism: A Very Short Introduction; Urbinati, Me the People.
 Müller, What Is Populism?, 3.
 Ibid., 19-20.
 Andrew Arato and Jean L. Cohen, “Civil Society, Populism, and Religion,” in Routledge Handbook of Global Populism, ed. Carlos de la Torre (New York: Routledge, 2018).
 Pierre Rosanvallon, Counter-Democracy: Politics in an Age of Distrust (Cambridge and New York: Cambridge University Press, 2008), 266.
 Paul Taggart, Populism (Buckingham and Philadelphia: Open University Press, 2000); Ernesto Laclau, “Populism: What’s in a Name?” In Populism and the Mirror of Democracy, ed. Francisco Panizza (London-New York: Verso), 2005; Francisco Panizza, “Introduction: Populism and the Mirror of Democracy,” in Populism and the Mirror of Democracy, ed. Francisco Panizza (London-New York: Verso, 2005); Julio F. Carrión, “Democracy and Populism in the Andes: A Problematic Coexistence,” in Latin American Democracy: Emerging Reality or Endangered Species? Second edition, ed. Richard L. Millett, Jennifer S. Holmes, and Orlando J. Pérez (New York and London: Routledge, 2015).
 Panizza, “Introduction: Populism and the Mirror of Democracy,” 29.
 Abts and Rummens, “Populism versus Democracy,” 420.
 Panizza, “Introduction: Populism and the Mirror of Democracy,” 29.
 Arato and Cohen, Populism and Civil Society, 155; Gábor Halmai, “Is There Such Thing as ‘Populist Constitutionalism’? The Case of Hungary,” Fudan Journal of Humanities and Social Sciences 11, no. 3 (2018): 323-339, https://doi.org/10.1007/s40647-018-0211-5.
 Maxwell A. Cameron and Kenneth E. Sharpe, “Andean Left Turns: Constituent Power and Constitution Making,” in Latin America’s Left Turns. Politics, Policies and Trajectories of Change, ed. Maxwell A. Cameron and Eric Hershberg (Boulder, CO, and London: Lynne Rienner Publishers, 2010), 65.
 Luigi Corrias, “Populism in Constitutional Key: Constituent Power, Popular Power, Popular Sovereignty and Constitutional Identity,” European Constitutional Law Review 12, no. 1 (2016): 6-26, quoted in Halmai, “Is There Such Thing as ‘Populist Constitutionalism’?,” 326, https://doi.org/10.1017/s1574019616000031.
 Javier Corrales, Fixing Democracy: Why Constitutional Change Often Fails to Enhance Democracy in Latin America (Oxford and New York: Oxford University Press, 2018), 10-11.
 Galston, Anti-Pluralism, 17.
 Javier Corrales, “Autocratic Legalism in Venezuela,” Journal of Democracy 26, no. 2 (2015): 37-51, https://doi.org/10.1353/jod.2015.0031.
 Kurt Weyland, “Latin America’s Authoritarian Drift: The Threat from the Populist Left,” Journal of Democracy 24, no. 3 (2013): 18-32, https://doi.org/10.1353/jod.2013.0045.
 David Landau and Rosalind Dixon, “Abusive Judicial Review: Courts against Democracy,” UC Davis Law Review 53, no. 3 (2020): 1313-1388, https://doi.org/10.1093/oso/9780192893765.003.0005.
 Carrión, A Dynamic Theory of Populism in Power, 225.
 Jan-Werner Müller, “Populism and Constitutionalism,” in The Oxford Handbook of Populism, ed. Cristóbal Rovira Kaltwasser, Paul Taggart, Paulina Ochoa Espejo, and Pierre Ostiguy (Oxford and New York: Oxford University Press, 2017), 591
 Steven Levitsky and James Loxton, “Populism and Competitive Authoritarianism in Latin America,” in Routledge Handbook of Global Populism, ed. Carlos de la Torre (New York: Routledge, 2018); Pappas, Populism and Liberal Democracy; Weyland, Populism’s Threat to Democracy; Carrión, A Dynamic Theory of Populism in Power.
 Catherine M. Conaghan, Fujimori’s Peru: Deception in the Public Sphere (Pittsburgh, PA: University of Pittsburgh Press, 2005); Levitsky and Loxton, “Populism and Competitive Authoritarianism in Latin America”; Weyland, Populism’s Threat to Democracy; Carrión, A Dynamic Theory of Populism in Power.
 Kurt Weyland, “How Populism Dies: Political Weaknesses of Personalistic Plebiscitarian Leadership,” Political Science Quarterly 137, no. 1 (2022): 9-42, 13, https://doi.org/10.1002/polq.13277.
 Santiago Basabe-Serrano and Julián Martínez, “Ecuador: Cada vez menos democracia, cada vez más autoritarismo…con elecciones,” Revista de Ciencia Política 34, no. 1 (2014): 145-170, https://doi.org/10.4067/s0718-090×2014000100007.; Carlos de la Torre, “The People, Democracy, and Authoritarianism in Rafael Correa’s Ecuador,” Constellations 21, no. 4 (2014): 457-466; Levitsky and Loxton, “Populism and Competitive Authoritarianism in Latin America,” https://doi.org/10.1111/1467-8675.12117.
 Conaghan, Fujimori’s Peru; Julio F. Carrión, ed., The Fujimori Legacy: The Rise of Electoral Authoritarianism in Peru (University Park, PA: The Pennsylvania State University Press, 2006).
 Fabrice Lehoucq, “Bolivia’s Citizen Revolt,” Journal of Democracy 31, no. 3 (2020): 130-144, https://doi.org/10.1353/jod.2020.0050.
 Steven Levitsky and Lucan Way, “The New Competitive Authoritarianism,” Journal of Democracy 31, no. 1 (2020): 51-65, https://doi.org/10.1353/jod.2020.0004.
 Eduardo Posada-Carbó, “Latin America: Colombia after Uribe,” Journal of Democracy 22, no. 1 (2011): 137-151.
 Maxwell A. Cameron, “Self-Coups: Peru, Guatemala, and Russia,” Journal of Democracy 9, no. 1 (1998): 125-139, https://doi.org/10.1353/jod.1998.0003.
 J. Mark Ruhl, “Honduras: Democracy in Peril,” in Latin American Politics and Development. Ninth edition, ed. Harvey F. Kline, Christine J. Wade, and Howard J. Wiarda (New York: Westview Press, 2018).
In Honduras, populist Xiomara Alfaro was elected in 2021, but she did not assume office until January 1, 2022. Therefore, she is not included in the analysis.
 Levitsky and Way, “The New Competitive Authoritarianism.”
 Orlando J. Pérez, “Panama: Political Culture and the Struggle to Build Democracy,” in Latin American Politics and Development. Ninth edition, ed. Harvey F. Kline, Christine J. Wade, and Howard J. Wiarda (New York: Westview Press, 2018).
 Morlino, “What is a ‘Good’ Democracy?”; Merkel and Croissant, “Conclusion: Good and Defective Democracies.”
 Andreas Schedler, “The Menu of Manipulation,” Journal of Democracy 13, no. 2 (2002): 36-50, https://doi.org/10.1353/jod.2002.0031.; Fabrice Lehoucq, “Electoral Fraud: Causes, Types, and Consequences,” Annual Review of Political Science 6, no. 1 (2003): 233-256, https://doi.org/10.1146/annurev.polisci.6.121901.085655.; Sarah Birch, Electoral Malpractice (Oxford: Oxford University Press, 2011).
 The index is built based on the indicators for EMB autonomy (v2elembaut), EMB capacity (v2elembcap), election voter registry (v2elrgstry), election vote-buying (v2elvotbuy), election other voting irregularities (v2elirreg), election government intimidation (v2elintim), non-state electoral violence (v2elpeace), and election free and fair (v2elfrfair) (Michael Coppedge et al., V-Dem Codebook v12. Varieties of Democracy (V-Dem) Project, 2022, 69).
 Ibid., 291.
 Christian Houle and Paul D. Kenny, “The Political and Economic Consequences of Populist Rule in Latin America,” Government and Opposition 53, no. 2 (2016): 256-287, https://doi.org/10.1017/gov.2016.25.; Huber and Schimpf, “Friend or Foe?”; Ruth-Lovell, Lührman, and Grahn, “Democracy and Populism, https://doi.org/10.1111/1467-9248.12219.
 Ruth, “Populism and the Erosion of Horizontal Accountability in Latin America;” Weyland, Populism’s Threat to Democracy.
 See Carrión, A Dynamic Theory of Populism in Power; Weyland, “How Populism Dies.”
 See, for example, Pappas, Populism and Liberal Democracy; Carrión, A Dynamic Theory of Populism in Power.
 See, for example, Weyland, Populism’s Threat to Democracy.
 Coppedge et al., V-Dem Codebook v12, 47.
 Cristóbal Rovira Kaltwasser and Paul Taggart, “Dealing with Populists in Government: A Framework for Analysis,” Democratization 23, no. 2 (2016): 201-220, https://doi.org/10.1080/13510347.2015.1058785.; Pappas, Populism and Liberal Democracy; Murat Somer, Jennifer L. McCoy, and Russell E. Luke, “Pernicious Polarization, Autocratization, and Opposition Strategies,” Democratization 28, no. 5 (2021): 929-948, https://doi.org/10.1080/13510347.2020.1865316.
 For a view that argues that polarization is a consequence of populism, see Kenneth M. Roberts, “Populism and Polarization in Comparative Perspective: Constitutive, Spatial and Institutional Dimensions,” Government and Opposition (2021): 1-23, https://doi.org/10.1017/gov.2021.14.
 Coppedge et al., V-Dem Codebook v12, 96.
 To test whether the main results were influenced by the decision to use fixed effects, I re-run Model 4 using random effects. The results were similar, with one exception: populism in power emerged as having a positive and statistically significant effect (0.069 at p=0.03) on CEI scores. The other findings were confirmed: horizontal accountability having very strong positive effects (at 0.001 or lower), and its interaction term with populism in power emerging with a negative sign, as expected, and achieving statistical significance at 0.001 or better.
 Rummens, “Populism as a Threat to Liberal Democracy”; Panizza, “Introduction: Populism and the Mirror of Democracy.”
 Urbinati, Me the People; Arato and Cohen, Populism and Civil Society; Müller, What Is Populism?