in ,

The Populist Harm to Democracy: An Empirical Assessment

The Populist Harm to Democracy: An Empirical Assessment

Jordan Kyle (Senior Fellow and Chief of Data and Analytics, Renewing the Centre) & Yascha Mounk (Executive Director, Renewing the Centre), Tony Blair Institut for Global Change, 26th December 2018

EXECUTIVE SUMMARY

Many academics and commentators are sounding the alarm about the threat that rising populism poses to the stability of liberal democracies. Others respond that populism is, on the contrary, a sign of democratic resilience, providing a necessary corrective that will help address popular grievances, curtail the excessive power of elites and make political systems more democratic.

To resolve this important debate on the basis of sound empirical evidence, this paper measures the impact that past populist governments have had on democracy by drawing on a first-of-its-kind global database of populist rule. It looks at the effect of populist government on three aspects of democratic institutions: the quality of democracy broadly, checks and balances on executive power, and political participation. The paper finds that populist rule—whether from the right or the left—has a highly negative effect on political systems and leads to a significant risk of democratic erosion.

KEY FINDINGS

  • Populists last longer in office. On average, populist leaders stay in office twice as long as democratically elected leaders who are not populist. Populists are also nearly five times more likely than non-populists to survive in office for over ten years.
  • Populists often leave office in dramatic circumstances. Only 34 per cent of populist leaders leave office after free and fair elections or because they respect term limits. A much larger number are forced to resign or are impeached, or do not leave office at all.
  • Populists are far more likely to damage democracy. Overall, 23 per cent of populists cause significant democratic backsliding, compared with 6 per cent of non-populist democratically elected leaders. In other words, populist governments are about four times more likely than non-populist ones to harm democratic institutions.
  • Populists frequently erode checks and balances on the executive.Over 50 per cent of populist leaders amend or rewrite their countries’ constitutions, and many of these changes extend term limits or weaken checks on executive power. The evidence also suggests that populists’ attacks on the rule of law open the way to greater corruption: 40 per cent of populist leaders are indicted on corruption charges, and the countries they lead experience significant drops in international corruption rankings.
  • Populists attack individual rights. Under populist rule, freedom of the press falls by some 7 per cent, civil liberties by 8 per cent and political rights by 13 per cent.

INTRODUCTION

Political scientists, journalists and even many politicians are more worried now about the stability of liberal democracy than they have been in many decades. According to the literature on the ‘crisis of democracy’, the world is in the midst of a democratic recession that threatens to set the cause of liberty and self-determination back by many decades; 1 is witnessing in real time what it looks like when seemingly stable democracies go through the process of decay or decomposition; 2 and might even face the threat of the rise of a new kind of fascism. 3

There are indeed some real reasons for concern. For the past 12 years, more countries have moved away from liberal democracy than towards it. Across the West, populist newcomers have taken on tremendous power in an astoundingly short span of time. 4 And in many of the countries in which populists have gained power, from Hungary to Turkey, they do seem to be inflicting serious damage on the institutions democracies need to sustain themselves. 5 The standard narrative—that the rise of authoritarian populism poses a new and especially dangerous challenge to the stability of liberal democracy—is rooted in a lot of striking contemporary evidence.

At the same time, there are also important reasons, both empirical and theoretical, to be sceptical of the crisis of democracy. Empirically, it is not clear that the rise of populism has, so far, significantly set democracy back either over the long run or on a global scale. Despite the democratic recession, for example, the overall number of countries that are democracies remains near an all-time high: in 2007, often marked as the year when democracy was at its highest globally, Freedom House ranked 48 per cent of countries as “free”; in 2017, Freedom House still ranked 45 per cent of countries as “free”. 6

Indeed, it is possible to cast the negative developments of the past years as an inevitable correlate of the earlier expansion of democracy. In the terms of political scientist Samuel Huntington, democracy rapidly expanded around the world during its “third wave” between the mid-1970s and 1990s. 7 The recent setbacks, then, are simply to be understood as a reverse wave, in which countries that had never fully consolidated struggle to hold onto their fragile democratic institutions.

The theoretical challenge goes even deeper. Perhaps populism is a feature rather than a bug—a sign of ongoing democratic resilience rather than democratic decay? According to this perspective, liberal democracies have always remained incomplete. Until recently, they have excluded many minority groups from full participation in the political and economic system. They placed far too much power in the hands of a narrow economic and social elite. And they designed an economic system that allowed the vast majority of the economic gains generated by globalisation to be captured by a small number of billionaires and giant corporations. Populism, activists like Stephen Bannon and theorists like Chantal Mouffe argue, is a necessary correction to these problems: over the long run, it will help address popular grievances, curtail the outsize power of political and financial elites, and even make political systems more fully democratic. 8

Explicitly or implicitly, both sets of theories rest on empirical assumptions. According to the people who believe that democracy is in crisis, populist governments often lead to a significant and lasting deterioration of democratic institutions. According to those who believe that this moment is far less dramatic, or might even present a big opportunity to deepen democracy, populism usually has a positive impact on the political system, by opening up new avenues for political participation. 9

This paper seeks to provide an empirical basis for these debates by drawing on a new, global database of populist rule. 10 Building on recent theoretical and empirical research, it assesses the impact of government by a populist leader on a country’s democratic system. The findings are as striking as they are worrying. Populist rule—whether from the right or from the left—has a highly negative impact on political systems and leads to a significant risk of democratic erosion.

At first sight, populist leaders are so different from each other that it seems to make little sense to group them all together. Venezuela’s late President Hugo Chávez, for example, claimed to stand up for traditionally disadvantaged groups and promised to expand the welfare state and take the fight to multinational corporations. Brazil’s President-Elect Jair Bolsonaro, by contrast, promises to curtail the welfare state and attract more investment to the country, and has quickly become infamous for a series of sexist, homophobic and racist slurs. Even the identities of the minority groups targeted by populists can differ wildly, both from movement to movement and from country to country: while populist leaders like France’s Marine Le Pen have historically targeted Muslims, others, like Turkey’s Recep Tayyip Erdoğan, have targeted non-Muslim minorities.

Despite these evident dissimilarities, populist movements share a striking number of attributes, including a disdain for elites and a profound distrust of establishment institutions. The first report in this series reviewed the extensive literature that addresses how to define populism and concluded that populists are united by two fundamental claims:

  1. Elites and ‘outsiders’ are working against the interests of the ‘true people’.
  2. Populists are the voice of the ‘true people’ of a country and nothing should stand in their way. 11

This conceptualisation of populism captures both its anti-elite orientation and its distinctive mode of political organisation, which involves bulldozing over political and civil-society institutions in the name of enacting the popular will. Using this definition, we have built a first-of-its-kind global database of populism, identifying 46 populist leaders or political parties that held executive office across 33 democratic countries between 1990 and 2018.

This data set allows us to give an empirical answer to one of the most urgent questions about the rise of populists: what is their effect on political institutions? The key assumption of political scientists who believe they have identified a threat to democracy is that populists tend to erode the rules and norms of existing political systems to such an extent that they do real damage to liberal democracy. By contrast, the central assumption of political scientists who believe that the current threat to democracy is overblown, or that populists may in fact have a salutary effect on the political system, is that populists help address popular grievances, rein in undemocratic elites, open up political participation to previously marginalised groups and thereby strengthen democracy. Which of these contrasting hypotheses is borne out by the data?

Comparative research on the effects of populist rule has only just begun, in part because until recently, populism was primarily confined to opposition movements. As a result, there is still no global, systematic understanding of what happens when populists gain public office in a broad range of cases and a broad range of regimes, for example in both presidential and parliamentary systems and across regions. 12 To remedy this shortcoming, this paper takes a wide view both in its global perspective on populist rule and in its assessment of the ways it might shape democratic institutions. The paper looks at the effect of populist rule on three aspects of democratic institutions: the quality of democracy broadly, checks and balances on executive power and political participation.

POPULISM’S EFFECT ON DEMOCRACY

In the past, democracies have typically broken down at gunpoint. Between 1960 and 1990, nearly three out of four democratic collapses were caused by coups d’état or the outbreak of civil war. 13 Argentina, Brazil, Nigeria, Pakistan and Turkey, among others, all suffered democratic breakdown through the use of violence, coercion and power by non-elected forces.

Of late, by contrast, democracies have increasingly started to break down because of the actions of democratically elected leaders who harness discontent with the functioning of democratic institutions to dismantle traditional limits on their power. Since the end of the Cold War, this has become the predominant form of backsliding, with a little over half of all cases propelled by democratically elected leaders. 14 This trend towards democratic collapse from within is even more prominent when the poorest and most unstable countries are excluded. Among countries with at least $1,000 in per capita gross domestic product (GDP), four in five democratic breakdowns since the end of the Cold War have been initiated from within. Among these, nearly two-thirds were brought about by populist leaders. 15

As political scientists Steven Levitsky and Daniel Ziblatt have pointed out, when democracies die from within, as they now usually do, they do so “slowly, in barely visible steps”. 16 Because these tactics are incremental rather than revolutionary, and usually take many years to complete, it is much more difficult today to pinpoint when a democracy dies than in the days when guns and tanks heralded democratic death. This difficulty is compounded by the fact that populists do not hide their intention to transform the political system. Promising to deliver more ‘wins’ for the ‘true people’, they discredit those who would oppose these tactics as part of an illegitimate cartel of elites or the complaints of a bitter opposition that has failed to win at the ballot box. By celebrating the expansion of executive power and the restriction of political competition as a democratic challenge to entrenched elites, they often manage to obscure the extent to which these changes empower populist leaders to override the will of the people if it turns against them in the future. 17

POPULISTS LAST LONGER IN OFFICE

The first and most basic mark of a liberal democracy is whether leaders respect electoral outcomes and, when they lose, whether they leave office through free and fair elections. This is why data on the longevity of populist governments are particularly important. The first striking finding is that contrary to the popular narrative, populists tend to stay in government for much longer than non-populists.

Many political scientists, historians and journalists believe that populists’ lack of political experience makes it difficult for them to sustain themselves in office. When Donald Trump was elected president of the United States in November 2016, for example, it was commonly predicted that he would not last a full term. Similarly, when two populist parties, the League and the Five Star Movement, formed an ideologically heterogeneous coalition in Italy in May 2018, many experts predicted that their marriage of convenience would quickly break down. But Trump is still in office, and his administration has proved more effective at instituting political change than many expected. Similarly, the Italian coalition, for now, shows no signs of breaking apart.

According to the data, this is more typical than atypical. In fact, populist governments tend to stay in office significantly longer than non-populists do. Using the Archigos database of political leaders, which identifies the effective leader of every country as well as leadership turnovers, we looked at the average length of time that populist leaders remain in power in a democracy, compared with non-populists. 18 The result is striking: populists, on average, hold office for twice as long as non-populists (six and a half versus three years).

A more formal way of examining whether populists stay in power for longer is to estimate the effect of populism on the risk of a leader leaving office during a given period. Figure 1 displays the results of such an analysis, showing the proportion of leaders still in office (vertical axis) after a given number of years in office (horizontal axis). The blue line represents the proportion of populist leaders surviving, while the grey line represents non-populists. As figure 1 shows, the probability of a populist remaining in office is consistently higher than the probability that a non-populist remains in office. 19

Figure 1: Proportion of Populists vs. Non-Populists Surviving in Office

A more intuitive way to visualise this is to ask how many times more likely a populist is to remain in power than a non-populist at any given time (see figure 2). After three years in office, only about half of non-populist governments survive. At the same juncture, the survival rate for populist governments is already significantly higher: 80 per cent of them are still in office. The ‘populist bonus’ becomes even more significant as governments are in office for longer. Ten years after they were first elected, only 6 per cent of non-populist governments are still in office. Populist governments, by contrast, retain more than a 1-in-4 chance of continuing to run their country. In other words, a populist leader is nearly five times more likely than a non-populist to survive in office for over ten years.

Figure 2: Populists’ Relative Likelihood of Staying in Office, Compared With Non-Populists

(Institutional features such as whether a country has a presidential or parliamentary system and background factors such as a country’s per capita income and other social or economic conditions also affect each leader’s predicted length of survival in office. In the appendix, we show that populist leaders are significantly more likely to stay in office than non-populist ones are, even after controlling for such country characteristics.)

POPULISTS OFTEN LEAVE OFFICE IN DRAMATIC CIRCUMSTANCES

It is not, in and of itself, a sign of democratic decline that a freely elected leader stays in office for an extended period. After all, a leader’s longevity in office might simply indicate the extent of his or her success and popularity. To figure out whether populists stay in office so much longer for benign or concerning reasons, it is therefore necessary to look at whether they ultimately leave office due to free and fair elections. Here, too, the finding is concerning: populist governments are not just more likely to stay in office for a long time; they also have a significant likelihood of leaving office under dramatic circumstances.

Out of the 47 times that a populist leader assumed office between 1990 and 2014, in only eight cases (17 per cent) did the leader step down after losing free and fair elections (see figure 3). 20 The same number stood down after reaching their term limit. A much larger number were forced to resign or were impeached. Populist presidents and prime ministers left office under such dramatic circumstances in 11 cases (23 per cent of the total). An even more common outcome—in 14 cases (30 per cent)—is that the populists have not left office at all and remain in power.

Figure 3: The Fates of Populists Who Entered Office Between 1990 and 2014

These findings raise an important question about the likely fate of the 30 per cent of populist governments in our database that remain in office. Are they more or less likely than the rest of the sample to leave office through free and fair elections? There is reason to fear that they are less likely to do so: the populist leaders who assumed power between 1990 and 2014 and remained in office as of late 2018 have already been in power for an average of nine years. Many are in countries that have undergone significant democratic erosion, such as Russia under President Vladimir Putin, Turkey under Erdoğan, Belarus under President Alexander Lukashenko, Venezuela under President Nicolás Maduro and Hungary under Prime Minister Viktor Orbán. Therefore, there is strong reason to worry that the populists in office today are even more likely than populist leaders as a whole to leave power in irregular and dramatic circumstances.

POPULISTS ARE FAR MORE LIKELY TO DAMAGE DEMOCRACY

The ample literature on democratic stability established long ago that political institutions are ‘sticky’, or resistant to change. In fact, the best predictor of whether a country is a democracy in any one year is whether it was a democracy the previous year. As a result, it is very unlikely that a country experiences democratic backsliding in any particular year. Of all the countries in our sample, only 5 per cent of democracies saw their democratic institutions erode in any one year. 21

But while institutions tend to be sticky, there are some prominent recent examples of democratic backsliding heralded by populist governments. The Polity IV database measures the extent to which a country is democratic on a scale from -10 to 10, with anything below 6 no longer considered a full democracy. 22 When Chávez assumed office in Venezuela in 1998, the country held a score of 8; by the time he died in office in 2013, it had fallen to -4. The demise of Turkish democracy has been even more dramatic: as recently as 2013, a full ten years into Erdoğan’s rule, it still scored 9; since then, it has fallen to -4.

How typical are these cases? And might there also be a significant number of non-populist governments that are responsible for democratic backsliding? The good news is that the likelihood of a democracy eroding under an elected non-populist leader is extremely low. Even under a reasonably broad definition of democratic backsliding, which includes any country in which the Polity score declined by a single point, only 6 per cent of non-populist leaders who assumed office after 1990 in a democratic country are responsible for this kind of deterioration.

The news about populist governments, by contrast, is far more alarming. While most countries survive populist governments without experiencing democratic backsliding, it is incontrovertible that they pose a severely heightened risk to the survival of democratic institutions. Overall, 24 per cent of populist leaders who assume office in a democratic country initiate democratic backsliding. In other words, a populist government is four times more likely than a non-populist one to damage democratic institutions. (It is likely that this undercounts actual cases of democratic erosion because of a status quo bias by the organisations that measure the robustness of democracies. Despite ample scholarly work demonstrating the erosion of the rule of law and media freedom in Hungary, for example, Polity IV had not yet registered democratic backsliding in the country as of 2017.)

These findings can shed light on another important question. According to commentators like John Judis, it is necessary to draw a sharp distinction between the effects of right-wing and left-wing populist governments. 23 In this view, the primary ideology unifying populist right-wing governments is nativism, an ideology that, in the words of political scientist Cas Mudde, “holds that states should be inhabited exclusively by members of the native group (‘the nation’) and that non-native elements (persons and ideas) are fundamentally threatening”; 24 left-wing populism, meanwhile, typically unifies around a more socially inclusive ideology promising to redress the failures of capitalism. Inclusionary left-wing populism, some commentators argue, is thus an antidote to right-wing populism, able to address voters’ anger and grievances without the baggage of exclusionary politics. 25

However, our data show that this argument does not bear out empirically. Between 1990 and 2014, 13 right-wing populist governments were elected; of these, five have significantly curtailed civil liberties and political rights, as measured by Freedom House. Over the same period, 15 left-wing populist governments were elected; of these, the same number reduced such freedoms. (Over the same period, there were also 17 populist governments that cannot easily be classified as either right- or left-wing; again, five of these governments diminished civil liberties and political rights.) Although this indicates a slightly higher rate of backsliding among right-wing populists than left-wing ones (38 per cent vs. 33 per cent), these data clearly contradict the belief that left-wing populism does not pose a threat to democracy.

POPULISTS ERODE CHECKS AND BALANCES ON EXECUTIVE POWER

Populist leaders often claim that they are exclusively beholden to the will of the people, making it legitimate for them to override legal or constitutional constraints on their power. 26 This basic claim can take very different forms. Left-wing populists tend to argue that the existing legal order only sustains the wealth and power of elites. Right-wing populists maintain that the existing rules and norms of democracy entrench the interests of snooty intellectuals or civil servants and convey illegitimate advantages on ethnic or religious outsiders. What both sets of claims have in common is that they empower populist governments to undermine checks on their authority, such as an independent judiciary. But do populists actually manage to weaken checks and balances on their power when they are in office?

One of the most basic indicators of whether populists are diminishing checks and balances on the executive is whether these leaders amend or rewrite their countries’ constitutions to increase the dependence of other branches of government on the executive and make it more likely that they will stay in power.

The result is deeply concerning: an astonishing 50 per cent of populists have changed or amended constitutions while in office, and many of these changes have extended executive term limits or weakened checks on executive power. 27

The extent to which populists damage the rule of law can helpfully be divided into two separate questions: First, do populists manage to destroy the rule of law as a whole, undermining the extent to which ordinary citizens can trust that justice is administered in a fair manner? And second, do populists manage to put themselves above the law, using their power to escape punishment for crimes or engage in large-scale corruption?

To answer the first question, we used the rule-of-law indicator in the World Bank Governance Indicators, which captures the extent to which members of society have confidence in, and abide by, the rules of society. In particular, it looks at the quality of contract enforcement, property rights, the police and the courts. 28 As predicted, populist rule is associated with a deterioration in the rule of law on this metric (see table 8 in the appendix). However, the size of the effect is small and not robust in all models.

At the same time, there is a much stronger indication that populists themselves flout the rule of law. Using our database on populist rule, we calculated the share of populists who assumed executive office between 1990 and 2014 and are indicted on corruption charges. Astoundingly, four in ten populist leaders are indicted on such charges. What is more, because only those populist leaders who do not erode an independent judiciary are ever charged in the first place, this likely understates the full extent of populist corruption. Forty per cent should therefore be understood as a conservative estimate of the share of populist leaders who personally profit from their hold on power.

Another indication that populists do little to rein in corruption while in power is that populists have led their countries to drop by an average of five places on Transparency International’s Corruption Perceptions Index, which ranks every country in the world based on levels of corruption. However, this average masks considerable variation. Some countries are ranked so poorly when they elect populists that there is not far to fall in the first place. Among the populist governments that start off ranked in the top half of the world, countries fall an average of 11 places under populist rule. Some cases are far more extreme than that: Venezuela, for example, dropped an astounding 83 places under the leadership of Chávez.

There is something bitterly ironic about the extent to which populists use their power for corrupt purposes. Discontent with corruption often brings populists to power. But they frequently end up even more corrupt than previous governments. In the words of economist Barry Eichengreen, voters who allow a leader to sweep away independent institutions “empower him to repopulate the swamp rather than draining it – to simply replace the mainstream’s alligators with his own”. 29

POLITICAL RIGHTS AND CIVIL LIBERTIES FALL UNDER POPULIST RULE

Populists seek to mobilise people against established power structures. 30 So there is reason to believe that populist movements might succeed in motivating masses of formerly disengaged citizens to participate in formal political structures. 31 At the same time, it is also possible that the negative rhetoric that populists introduce into politics might, over the long run, deepen disenchantment and distrust towards politics, parties and democracy, leading more citizens to check out of formal politics altogether.

One way to assess whether populists enhance or deter political participation might be to examine whether populist candidates affect voter turnout. But this can be difficult to interpret. Populists might mobilise new voters, yet they have strong incentives to deter political participation among their opponents. 32 Higher voter participation, conversely, could just as easily stem from a deep fear that the stakes of politics have become too high as from enthusiasm about populist policies. Thus, the net effect on voter turnout could be zero, even if who votes is changing. Indeed, studies examining the effect of populist parties on voter turnout tend to find no overall increase. 33

What is more, even if populists did increase voter participation, it would be difficult to interpret what this means. Lower voter participation could indicate that citizens are deeply disenchanted with politics, that they feel certain one side will win and their vote will not be pivotal, or that they feel comfortable enough with any electoral outcome that it is not vital to vote. Even citizens living under the most authoritarian regimes still vote; yet, without the right to express themselves freely, organise around their preferences and choose between political candidates, their votes mean little.

Instead of examining whether more citizens participate in politics by turning out on election day, then, we considered whether citizens have the right to participate in a meaningful way. At least three conditions are necessary for this:

  • A free press offers citizens the opportunity to make informed choices and hold leaders accountable.
  • Civil liberties ensure that citizens have freedom of expression and association—freedoms that enable citizens to voice their preferences and beliefs and to organise around their interests and values.
  • Political rights ensure that almost all adult citizens have the right to vote, that they can participate in free and fair elections, and that the opposition plays a real and important role in such elections.

Replicating the models used throughout this paper, we found that populist rule is associated with less freedom of the press, fewer civil liberties and fewer political rights (see tables 9, 10 and 11 in the appendix). We measured each of these concepts using data from Freedom House, which makes annual assessments of countries’ freedom of the press, civil liberties and political rights. Populist rule is associated with a 7 per cent decline in press freedom, an 8 per cent fall in civil liberties and a 13 per cent decrease in political rights. Considering that these changes hold constant a country’s history, time shocks, per capita income and many other fundamental drivers of democratic quality, their magnitude is astounding.

CONCLUSION

The findings from our global database on populist governments, and these governments’ impact on democratic persistence, are alarming. Political scientists who have emphasised the danger populists pose to the persistence of democratic institutions are vindicated by the historical record: unfortunately, there really is a strong empirical link between the rise of populism and an increase in democratic backsliding. But especially because these findings underline the tension between populism and democracy, it is also important to point out what they do not imply.

First, many critics of the growing literature on populism fault mainstream accounts of populism for seeking to delegitimise popular grievances. On this view, scholars who point out that populism often undermines democracy are too unwilling to acknowledge that populist voters are often motivated by perfectly legitimate concerns about their countries’ shortcomings. But this sets up a false binary. It is perfectly possible both to recognise the dangers populism poses to democracy and to believe that the willingness of a growing number of citizens to jettison existing institutions has deep structural causes. An adequate defence of democracy is therefore likely to involve a double-edged strategy, which strives to undercut the support for populism by identifying and addressing these grievances as well as to defend democratic institutions by opposing populist forces that are out to destroy them.

Second, the historical record strongly suggests that populism is a clear and present danger to democracy. But while this finding is a reason for serious concern, it is no excuse for fatalism. For although populism is one of the principal causes of democratic death, most democracies that are faced with a populist government do manage to survive. It is therefore of the greatest importance to use the historical record to examine which opposition strategies have tended to succeed, and which have generally failed. This urgent task will be the subject of a follow-up publication.

So what are the lessons for some of the richest and most powerful democracies that are now governed by populist leaders? And do the findings presented here bode well or ill for the ability of American institutions to withstand Trump’s populist administration?

To answer these questions, it is important to understand that there are significant differences between the United States and virtually all of the other countries in our database. The United States is the oldest continuously functioning democracy in the world. It is one of the world’s richest democracies. And its political system is extremely decentralised, making it much harder for one government to concentrate power in its hands. All of these facts make it more difficult for American presidents to undermine democratic institutions. It would therefore be rash to conclude that populists will necessarily have as much success in dismantling democratic institutions in the United States as they already have in countries like Hungary or Venezuela.

But while there is no historical precedent that is sufficiently similar to allow us to estimate the effect of a populist government on the durability of political institutions in a rich, long-established democracy like the United States, it would be a mistake to assume that these dissimilarities neutralise any danger. Indeed, in his first years in office, the country’s current president has already pursued many of the same basic strategies that other populists around the world have effectively used to weaken their own democratic institutions. And although the United States remains a functioning democracy today, many of its institutions have already proved more susceptible to pressure from the executive than optimists had predicted.

APPENDIX: METHODOLOGY

CODING POPULISM

The data in this paper rely on the database “Populists in Power, 1990–2018” developed by Jordan Kyle and Limor Gultchin. Full details of the methodology can be found in the methodological appendix to their report Populists in Power Around the World34 A summary is offered below.

The database supports efforts to build a systematic understanding of how populists govern, including how they reshape state institutions, how they may or may not erode the quality of liberal democracy, and the economic policies they implement. To understand these questions across a wide range of social, economic and political contexts, a global accounting of populism in power is necessary.

To make the project cross-regional, the focus of this project is on both leaders and parties that can be classified as populist. While parliamentary systems tend to give precedence to political parties, presidential systems favour individual leaders. This analysis focuses on populist parties and leaders who attained executive office in at least minimally democratic countries between 1990 and 2016. 35 This includes only those populists who reached the presidency or prime ministership (or the equivalent executive office), and not those who governed as minority partners in a coalition government. 36 Specifically, we used the Archigos database of political leaders, which identifies the effective leader of every country in every year going back to 1875.

Requiring that countries have attained a certain level of democracy to be included leaves off many instances of populism that have risen in semi-democratic or authoritarian settings. This omits, for example, many cases of African and Middle Eastern populism. Similarly, requiring that the populist leader or party has attained the highest executive office ignores many instances where populism has been highly influential yet has never risen to the level of controlling the executive branch. In this sense, the database conservatively undercounts the global incidence and influence of populism.

Classifying particular parties and leaders as populist is a fraught exercise, due to the many disagreements on the definition of populism and the fact that populism is hardly a binary phenomenon that is either fully present or fully absent. Some leaders may be readily identifiable as full-blown populists, yet many sit on the boundary. Moreover, to the extent that populism is a political strategy that can be adopted to different degrees by different actors over time (rather than a strict political doctrine that actors either subscribe to or not), the presence or absence of populism is a matter of degree that can vary over time.

Given the difficulty of this exercise, a reasonable place to start is the extensive scientific literature on populism and the deep well of subject matter and case-study expertise that can be found there. Even though the literature famously disagrees on the exact definition of populism, there is, according to political scientist Benjamin Moffitt, “at least some (mild) consensus regarding the actual cases of actors that are usually called ‘populist’”. 37 This can be seen in the fact that scholars of populism tend to reference the same set of cases over and over.

To identify leaders associated with populism, we developed a three-step process. First, we identified 66 leading academic journals in political science, sociology and area studies that commonly publish articles on populism, as well as the new Oxford Handbook of Populism. From these sources, we queried all articles containing the keyword “populist” or “populism” in their title or abstract and scanned the texts using natural language processing technology that can identify names. These names emerged as the potential list of populist leaders.

Second, from this potential list, we carefully read each source to ensure that we included only those with substantive discussion of why the leader in question qualified as populist. We reviewed the sources for each case to verify that the leader in question met both of the elements of the definition of populism set out in the report.

Third, we sent the list of potential populist leaders that emerged from this exercise to several populism experts, to verify both whether the leaders from their region of expertise met their understanding of populism and whether there were any additional leaders whom we may have missed. To investigate these additional leaders, we often reached beyond the initial list of leading academic journals and books to other peer-reviewed specialist journals and case-specific academic books. In short, for every potential case of populism that emerged either from our initial text searches or from our consultations with experts, we consulted as many credible sources as possible to ascertain whether the case in question met our definition of populism.

POPULISTS’ LENGTH OF TIME IN OFFICE

One indicator of whether populists are eroding democratic norms and institutions is whether they tend to stay in office longer than democratically elected non-populist elected leaders. Duration models estimate the effect of populism on the ‘risk’ that a leader leaves office during a given period, conditional on the length of his or her tenure until that point. If the leader did not leave office by 31 December 2015, the data are considered censored.

Table 4 reports hazard ratios from Cox proportional hazards models rather than coefficients for each independent variable. Hazard ratios are interpreted relative to 1: a hazard ratio greater than 1 means that high values of that variable increase the risk that a leader leaves office; hazard ratios less than 1 indicate variables that decrease the risk that a leader leaves office. 38 For example, a hazard ratio of 0.67 suggests that a one-unit change in the independent variable is associated with the risk of a leader losing office dropping by one-third.

Column 1 reports results from the simplest possible duration model, with no co-variates that vary with time. In this case, the dependent variable is the number of months that the leader spends in office, and per capita income and the number of years that a country has been a democracy are measured at the point in time that a leader enters office. 39 Populism is linked with a 57 per cent reduction in the risk that a leader leaves office in any given year.

Table 4: How Long Leaders Stay in Office, 1990–2014

Dependent VariableTime in Office (1)
Populist rule0.432 (0.094)**
Per capita income (log)1.046 (0.067)
Presidential system0.714 (0.148)
# years as a democracy (log)0.999 (0.004)
Civil conflict0.957 (0.142)
N527

** p < 0.01. Standard errors clustered by country in parentheses.

DEMOCRATIC BACKSLIDING

Democracies that elect populists are at a far greater risk of democratic backsliding than democracies that elect non-populist leaders. However, the types of democracies that elect populists may be less consolidated and more likely to backslide in the first place. The extent to which democracy has consolidated in a country is about much more than the level of democracy in that country today. Rather, we considered the overall stock of democracy in a country: a country’s history with democratic institutions. 40

We additionally controlled for other factors that may contribute to democratic backsliding in a country: per capita income (logged), whether a country has a presidential system, average growth rates during the past three years and whether the country has an ongoing civil conflict. The models include a lagged dependent variable (that is, the democracy score of the country in the previous year), so the model estimates the effect of populism on the change in the democracy score. We also include regional and year fixed effects to account for any systematic differences between regions or over time.

Given the importance of estimating changes in the democracy score—rather than the level—we also estimate country fixed effects models. Country fixed effects control for any background characteristics of particular countries that do not change over time. Rather than examining why countries have different democracy scores overall, then, the fixed effects model focuses on why democracy scores in individual countries change over time. Variables that do not vary in countries over time—such as whether the system is presidential or parliamentary—cannot be included in the fixed effects models, as the fixed effects already account for any factors that are constant over time.

Finally, we looked at models that consider whether the length of time that a populist holds office shapes the likelihood of backsliding. These models interact populist rule with the amount of time a leader has been in office. Coefficients from three different variables are reported: “populist rule” is equal to one in any year in which a populist is in power in a given country; “time in office” is equal to the number of years the leader of a country has been in power; and “# years populist has been in office” multiplies these two terms together, so it is equal to the zero if the leader is not a populist and equal to time in office if the leader is a populist.

The coefficient on “populist rule” in this case represents the effect of populist rule if the populist has been in power for zero years. Thus, it is a rather meaningless coefficient. The coefficient on “time in office” represents the relationship between the time a non-populist leader spends in office and the likelihood of democratic erosion. The coefficient on “# years populist has been in office” is the real coefficient of interest in these models, and can tell us how the likelihood of democratic backsliding evolves as the populist is in power for longer and longer.

Table 5: Effect of Populist Rule on Democratic Backsliding, 1990–2014

Dependent VariablePolity IV Democracy Score
(1)(2)(3)
Populist rule-0.153 (0.077)*-0.797 (0.429)†0.096 (0.290)
Leaders’ time in office-0.044 (0.013)**
# years populist has been in power-0.303 (0.088)**
Per capita income (log)0.047 (0.021)*1.696 (0.545)**2.221 (0.503)**
Growth rate (three-year average)0.018 (0.010)†-0.088 (0.050)†-0.064 (0.041)
Presidential system0.054 (0.035)
# years as a democracy (log)-0.015 (0.022)0.455 (0.175)*0.396 (0.146)**
Civil conflict-0.050 (0.067)-0.744 (0.502)-0.695 (0.463)
Lagged dependent variableYESNONO
Regional fixed effectsYESNONO
Year fixed effectsYESYESYES
Country fixed effectsNOYESYES
N202121292103

** p < 0.01; * p < 0.05; † p < 0.10. Standard errors clustered by country in parentheses.

Even controlling for many of the key ways in which countries with populist rule may differ from countries without populist rule, such rule is linked with democratic backsliding: populist rule is associated with a 10 per cent drop relative to the mean democracy score. Compare this with the effect of per capita income on democratic backsliding: a country would have to drop from $16,000 to $10,000 in per capita income (a fall of 38 per cent) to see a similar risk of democratic backsliding.

What is particularly striking about the democratic backsliding overseen by populists is the strong correlation between the length of their tenure and the extent to which they damage institutions. Indeed, the first four years that a populist is in office are associated with a 14 per cent drop in a country’s Polity score (relative to the sample mean), while being in office for eight years is associated with a 29 per cent drop in the Polity score. Since many democracies run on a four-year electoral cycle, this implies that the stakes for democratic survival increase each time populists seek to renew their mandate. While most countries can contain populists’ attacks on democratic institutions while they are first in office, the system’s capacity to keep populists’ authoritarian instincts in check is significantly weakened each time a populist wins re-election.

Another way to assess whether populism is leading to democratic backsliding is to use duration analysis. In this case, we evaluated whether populist rule affects the risk that a democracy breaks down altogether (that is, drops to a Polity score below 6). Table 6 reports the results from a Cox proportional hazard model, taking the time until a democracy breaks down as the dependent variable. Countries that retained democratic institutions until 31 December 2015 are considered censored. Note that this model includes time-varying co-variates. Populist rule is associated with a 13 per cent increase in the risk that a democracy breaks down in any given year.

Table 6: Populism and the Duration of Democracy, 1990–2014

Dependent VariableTime to Democratic Breakdown (1)
Populist rule1.132 (0.038)**
Per capita income (log)0.919 (0.032)*
Growth rate (three-year average)0.977 (0.020)
Presidential system0.479 (0.332)
# years as a democracy (log)0.971 (0.033)
Civil conflict1.118 (0.033)*
N1820

Standard errors clustered by country in parentheses.

POPULISM AND CHECKS AND BALANCES ON THE EXECUTIVE

We also tested specifically whether populists are more likely than non-populists to reduce institutional checks and balances on the executive. We measured checks and balances using the Constraints on the Executive variable from Polity IV, which measures the extent of institutionalised constraints on the decision-making powers of chief executives. In democracies, this involves having an independent, empowered legislature and a strong, independent judiciary.

Table 7 reports the results from the analysis. Controlling for per capita income (logged), whether a country has a presidential system, average growth rates during the past three years, the number of years that a country has been democratic (logged), whether the country has an ongoing civil conflict, regional and year fixed effects, and the country’s level of executive constraints in the previous year, we found that populist rule is associated with a 4 per cent drop in the country’s score on the executive-constraints indicator relative to the sample mean. A four-year term is associated with a 6 per cent drop in the country’s executive constraints indicator (relative to the sample mean).

Table 7: Effect of Populist Rule on Constraints on the Executive, 1990–2014

Dependent VariableConstraints on the Executive
(1)(2)(3)
Populist rule-0.053 (0.028)†-0.270 (0.202)-0.007 (0.149)
Leaders’ time in office-0.031 (0.013)*
# years populist has been in power-0.090 (0.038)*
Per capita income (log)0.018 (0.008)*0.553 (0.217)*0.638 (0.197)**
Growth rate (three-year average)0.009 (0.004)*-0.0270 (0.017)-0.021 (0.015)
Presidential system0.023 (0.016)
# years as a democracy (log)0.001 (0.010)0.166 (0.076)*0.189 (0.078)*
Civil conflict-0.026 (0.027)-0.287 (0.241)-0.280 (0.227)
Lagged dependent variableYESNONO
Regional fixed effectsYESNONO
Year fixed effectsYESYESYES
Country fixed effectsNOYESYES
N202121292103

** p < 0.01; * p < 0.05; † p < 0.10. Standard errors clustered by country in parentheses.

We also looked at whether populism affects rule of law in a country. We used the rule-of-law indicator of the World Bank Governance Indicators, which captures the extent to which agents have confidence in and abide by the rules of society. In particular, it looks at the quality of contract enforcement, property rights, the police and the courts. 41 Table 8 reports the results.

Table 8: Effect of Populist Rule on the Rule of Law, 1990–2014

Dependent VariableRule of Law
(1)(2)(3)
Populist rule-0.018 (0.007)**-0.059 (0.041)-0.047 (0.041)
Leaders’ time in office-0.004 (0.002)†
# years populist has been in power-0.003 (0.010)
Per capita income (log)0.0134 (0.005)**0.420 (0.175)*0.444 (0.175)*
Growth rate (three-year average)0.003 (0.001)*-0.010 (0.006)†-0.011 (0.006)†
Presidential system0.005 (0.005)
# years as a democracy (log)0.006 (0.004)0.029 (0.036)0.029 (0.036)
Civil conflict-0.007 (0.009)-0.042 (0.071)-0.040 (0.072)
Lagged dependent variableYESNONO
Regional fixed effectsYESNONO
Year fixed effectsYESYESYES
Country fixed effectsNOYESYES
N157316871668

** p < 0.01; * p < 0.05; † p < 0.10. Standard errors clustered by country in parentheses.

We found that, as predicted, populist rule is associated with a deterioration in the rule of law on this metric. However, the effect size is substantively small and not robust in all models.

POPULISM, POLITICAL RIGHTS AND CIVIL LIBERTIES

Replicating the same models with the same controls used throughout this paper, we found that populist rule is associated with less freedom of the press, fewer civil liberties and fewer political rights (see tables 9, 10 and 11). We measured each of these concepts using data from Freedom House, which makes annual assessments of a country’s press freedom, civil liberties and political rights. For each of these variables, we reversed the scales so that higher values indicate more freedom.

Populist rule is associated with a 7 per cent decline in freedom of the press, an 8 per cent fall in civil liberties and a 13 per cent drop in political rights (relative to sample means).

Table 9: Effect of Populist Rule on Press Freedom, 1990–2014

Dependent VariableFreedom House Press Freedom
(1)(2)(3)
Populist rule-0.945 (0.314)**-2.570 (1.477)†0.081 (1.572)
Leaders’ time in office-0.118 (0.094)
# years populist has been in power-0.953 (0.500)†
Per capita income (log)0.207 (0.099)*11.80 (3.348)**12.50 (3.093)**
Growth rate (three-year average)0.132 (0.028)**-0.397 (0.121)**-0.373 (0.114)**
Presidential system0.109 (0.171)
# years as a democracy (log)0.011 (0.122)1.085 (1.146)1.121 (1.182)
Civil conflict-0.307 (0.365)-6.408 (1.751)**-6.171 (1.681)**
Lagged dependent variableYESNONO
Regional fixed effectsYESNONO
Year fixed effectsYESYESYES
Country fixed effectsNOYESYES
N174818921870

** p < 0.01; * p < 0.05; † p < 0.10. Standard errors clustered by country in parentheses.

Table 10: Effect of Populist Rule on Civil Liberties, 1990–2014

Dependent VariableFreedom House Civil Liberties
(1)(2)(3)
Populist rule-0.071 (0.030)*-0.187 (0.093)*-0.059 (0.086)
Leaders’ time in office
# years populist has been in power
Per capita income (log)0.062 (0.012)**0.686 (0.242)**0.792 (0.224)**
Growth rate (three-year average)0.004 (0.004)-0.019 (0.011)-0.015 (0.010)
Presidential system0.0491 (0.016)**
# years as a democracy (log)0.004 (0.012)0.142 (0.075)†0.144 (0.074)†
Civil conflict-0.086 (0.037)*-0.150 (0.114)-0.148 (0.113)
Lagged dependent variableYESNONO
Regional fixed effectsYESNONO
Year fixed effectsYESYESYES
Country fixed effectsNOYESYES
N197020772052

** p < 0.01; * p < 0.05; † p < 0.10. Standard errors clustered by country in parentheses.

Table 11: Effect of Populist Rule on Political Rights, 1990–2014

Dependent VariableFreedom House Political Rights
(1)(2)(3)
Populist rule-0.085 (0.026)**-0.280 (0.151)†-0.030 (0.125)
Leaders’ time in office-0.015 (0.007)*
# years populist has been in power-0.084 (0.030)**
Per capita income (log)0.043 (0.011)**0.784 (0.379)*0.946 (0.351)**
Growth rate (three-year average)0.001 (0.004)-0.022 (0.012)†-0.015 (0.010)
Presidential system0.042 (0.019)*
# years as a democracy (log)0.008 (0.013)0.060 (0.090)0.045 (0.084)
Civil conflict-0.053 (0.033)-0.222 (0.212)-0.211 (0.199)
Lagged dependent variableYESNONO
Regional fixed effectsYESNONO
Year fixed effectsYESYESYES
Country fixed effectsNOYESYES
N197020772052

** p < 0.01; * p < 0.05; † p < 0.10. Standard errors clustered by country in parentheses.

ACKNOWLEDGEMENTS

We thank Louise Paulsen and Limor Gultchin for excellent research assistance.

N&R

Sobre as consequências nefastas do populismo

Há um movimento militar no Brasil atual: só um tolo não vê