The global rise of political polarization has fueled concerns about its detrimental impact on politics and society. From an increase in political violence to a decrease in the quality of democracy and governance, the threats posed by pernicious polarization—the division of society into two mutually antagonistic political camps—are diverse and acute. Determining how to reduce these tensions is therefore an urgent challenge. Using data on political polarization from the Varieties of Democracy (V-Dem) data set, this working paper assesses various instances of depolarization around the globe since 1900 and analyzes their long-term sustainability.
Polarization is increasing worldwide. When broken down by region, V-Dem data suggest that every region except Oceania has seen polarization levels rise since 2005. Africa has had the smallest increase during this period, although it has long had high levels of polarization. Rising polarization in Europe is being driven by deepening political divisions in Eastern and Central Europe, Southern Europe, and the Balkans. In the Western Hemisphere, the largest democracies—Brazil, Mexico, and the United States—are all experiencing extreme levels of polarization. East Asia’s polarization levels have traditionally been low, though increasing political tensions in places like South Korea and Taiwan are driving up the region’s score. And in South Asia, India’s polarization has skyrocketed since 2014.
To better understand the various paths by which polarized societies might overcome or reduce their political divisions, this working paper examines perniciously polarized countries that have successfully depolarized, at least for a time. Through a quantitative analysis of the V-Dem data set, this study identifies 105 episodes from 1900 to 2020 where countries were able to reduce polarization from pernicious levels for at least five years. These 105 episodes represent roughly half of the total episodes of pernicious polarization during the time period, thus indicating a fairly robust capacity of countries to depolarize. If considered in terms of country experiences rather than episodes (because many countries have experienced multiple episodes in a cycle of polarization and depolarization), then the data indicate that two-thirds of the 178 countries for which V-Dem provides polarization data have experienced one or more episodes of pernicious polarization, but only thirty-five countries (20 percent) have failed to experience any depolarization to below-pernicious levels.
Given this apparent capacity among the majority of the world’s countries to depolarize from pernicious levels at least some of the time, this analysis seeks to identify the contexts and sustainability of those experiences and to encourage further research into their causal mechanisms and outcomes for democracies. The analysis offers a preliminary discussion of the potential meaning and normative implications of depolarization as a concept and policy goal. It then uses qualitative analysis to identify patterns in the contexts of various depolarization cases and gauge the sustainability of these trends.
Most of these depolarization episodes were associated with dramatic changes in a country’s political life. An analysis of contextual factors showed that almost three-quarters of the cases came after major systemic shocks: a foreign intervention, independence struggle, violent conflict, or regime change (primarily in a democratizing direction). In the remainder of cases, countries depolarized within a given regime structure, whether democratic or autocratic. Tellingly, the authors identified no cases of depolarization from pernicious levels among liberal democracies. This is because very few countries classified as full liberal democracies have ever reached pernicious levels; the United States stands out today as the only wealthy Western democracy with persistent levels of pernicious polarization.
Just under half of all depolarizing cases were able to sustain depolarization for a decade or longer. In a second sizeable group of cases, countries managed polarization to some degree, either living with chronic near-pernicious levels after depolarizing, or else repolarizing to near-pernicious levels within ten years. Finally, 15 percent of the cases returned to pernicious levels within the first decade. Troublingly, when the authors analyzed the entire time period from 1900 to 2020, nearly half of the countries that had sustained depolarization or managed polarization for at least a decade later returned to pernicious levels of polarization.
These outcomes illustrate the difficulty of sustaining low levels of polarization, and they indicate that a cyclical pattern of polarization, depolarization, and repolarization may be a characteristic feature of political life in many places. Only a fraction (14 percent) of cases resulted in sustained depolarization over the long term.The mechanisms and strategies that enable such sustained depolarization in democracies will be the subject of future research. But given the small number of democracies (eleven) able to accomplish this feat amid the larger pattern of cyclical polarization and depolarization, it will be crucial to also understand strategies of managing polarization at moderately high levels while avoiding democratic erosion, government dysfunction, or returns to pernicious polarization and potential violence.
In the last decade, the division of societies into “us versus them” political camps has prompted alarm over the threats polarization poses to political life around the world. An emerging body of scholarly work points to extreme polarization as a key factor contributing to government dysfunction, political conflict, democratic erosion, and incremental autocratization.1 Thus, finding ways to depolarize societies or to manage polarization while ensuring governability is an urgent task.
To better understand the various paths by which societies might overcome or reduce political divisions, this working paper examines perniciously polarized countries that have successfully depolarized, at least for a time. It seeks to identify patterns in the political contexts of these countries and gauge the sustainability of these depolarization cases. To the authors’ knowledge, this study represents the first attempt to identify a comprehensive list of depolarization episodes and determine the contexts in which they occurred. Though it is an early effort, the 105 episodes provide a clear picture of where, when, and for how long depolarization has come about in recent history (see appendix A). It is worth noting at the outset that these patterns are descriptive and not explanatory. This data set will provide the basis for future research to investigate causal mechanisms and strategies used in various contexts.
The working paper is organized as follows. The first section defines polarization and depolarization, examines global trends in polarization, and describes how such trends affect democratic quality. The second section outlines the methods for case selection and variable measurements. The third section analyzes the contexts of each depolarization episode. In some cases, countries were weathering major political changes such as the aftermath of violent conflict, independence struggles, changes in their political regimes, or foreign interventions. In other cases, depolarization occurred within autocratic or democratic regimes without these other catalysts. The fourth section then examines the sustainability of these depolarization episodes, identifying which countries sustained this momentum, which ones simply managed near-pernicious levels of polarization, and which of them repolarized to pernicious levels over the subsequent decade. It also examines levels of polarization in these countries in the long run. Finally, the fifth section discusses lessons learned from these broad patterns of depolarization.
The Problem of Polarization and Its Consequences for Democracy
To assess the causes and consequences of polarization and depolarization, it is important to outline how these concepts are defined and measured.
Defining Polarization and Depolarization
Polarization can be thought of as a process, a state of equilibrium, and a political strategy. It is a process of simplifying politics in ways that lead toward a binary division of society into mutually antagonistic camps.2 Some degree of political polarization is natural and healthy in a democracy to distinguish the platforms of competing political parties and to encourage citizens to participate in politics more when major policy changes are needed. But as the process of polarization deepens and is prolonged, “the normal multiplicity of differences in a society increasingly align along a single dimension and people increasingly perceive and describe politics and society in terms of ‘Us’ versus ‘Them.’”3 Scholars thus increasingly analyze the growing partisan divides among citizens in psychological terms of social identity and intergroup conflict, in which each member becomes fiercely loyal to their side and wants it to win at all costs, while expressing strong bias or prejudice against the other group.4 Social identifiers such as ethnicity, social class, religion, language, or place of residence become aligned with one’s partisan identity.5
When polarization reaches a state of equilibrium, with a society divided into binary, mutually distrustful political camps where neither side has an incentive to pursue a depolarizing strategy, it has pernicious consequences for democracy: parties become unwilling to compromise, voters lose confidence in public institutions, and normative support for democracy may decline. In extreme cases, each camp begins to view the opposing camp and its policies as an existential threat to its own way of life or the nation as a whole. They come to perceive the “Other” in such negative terms that a normal political adversary competing for power is transformed into an enemy to be vanquished. This can be called pernicious polarization. It is different from lesser states of polarization in that the strength of animosity and distrust between the camps, its entrenchment in political dynamics, and its negative consequences for democracy are more severe. Pernicious polarization is difficult to reverse because of equilibrium conditions that incentivize polarization-reproducing behavior by all sides, and reversing it may take external shocks, major sociopolitical upheavals, or purposeful collective action and political strategies.6
Finally, polarization can also be used instrumentally by political elites as a strategy to gain and retain power.7 Postures of Manichaean moralizing judgment, identifying the ingroup as good and the outgroup as evil, are a nefarious aspect of this polarizing strategy, which aims to discredit the moral legitimacy of an opposition. Prior research has also found that polarization around “formative rifts”—unresolved historical debates around citizenship, national identity, and early myths from a country’s founding—has a particularly divisive quality because formative rifts cannot be eliminated without fundamentally reconfiguring the state in question.8 Because people often find themselves on one side of these rifts or the other by birth, activating these rifts is likely to be socially and psychologically divisive and to involve questions about who should be viewed as rightful citizens and who should represent them.
This definition of polarization is descriptive in the sense that it is helpful for identifying, measuring, and modeling polarization, and it is explanatory in the sense that it is useful for understanding and explaining the processes it entails and the outcomes it produces.
By contrast, no such concept of depolarization exists, given the lack of theoretical or conceptual work on the topic. This analysis offers an operational definition, viewing depolarization merely as a reduction in the level of political polarization as measured by experts’ assessments of the level of hostile interactions between political camps.
Although this operational definition of depolarization appears straightforward, understanding the limitations of this definition is crucial to correctly interpreting this study’s findings. First, no one yet knows the degree to which the social and political effects of polarization and depolarization are symmetrical—that is to say, whether the consequences of a certain amount of polarization can be reversed by the same amount of depolarization. The authors’ earlier research indicates that the incentives created by the logic of pernicious polarization make it hard to reverse. Significant evidence shows how a rise in the level of polarization affects society and politics, but comparable data on depolarization is lacking. Second, and relatedly, while polarization is a divisive process that simplifies the complexity of politics by emphasizing “us versus them” divisions, it remains unclear whether depolarization serves as a unifying and reconciling process in which cross-cutting ties reemerge. For example, depolarization may result from repression by autocratic regimes, but repression is unlikely to generate solidarity. In short, depolarization (like polarization) can look very different and have different mechanisms depending on context.
Until recently it was difficult to measure and compare pernicious polarization cross-nationally; instead, qualitative comparative studies and proxy measures dominated analysis of its mechanisms and effects.9 However, in 2020, the Varieties of Democracy (V-Dem) Institute released a new measure for political polarization, allowing for cross-national empirical studies.10 Grounded in survey data from over 3,000 country experts, V-Dem’s unique methodological approach also allows for longitudinal studies of data on political dynamics. With an average of five to seven experts assessing specific indicators for each country, the V-Dem data set covers 202 countries and political entities from 1900 to 2020. V-Dem’s political polarization metric asks country experts to answer the question: “To what extent is society divided into mutually antagonistic camps in which political differences affect social relationships beyond political discussions.”11 The measure uses a 0 to 4 scale with the following possible responses:
- 0: “Not at all. Supporters of opposing political camps generally interact in a friendly manner.”
- 1: “Mainly not. Supporters of opposing political camps are more likely to interact in a friendly than a hostile manner.”
- 2: “Somewhat. Supporters of opposing political camps are equally likely to interact in a friendly or hostile manner.”
- 3: “Yes, to a noticeable extent. Supporters of opposing political camps are more likely to interact in a hostile than friendly manner.”
- 4: “Yes, to a large extent. Supporters of opposing political camps generally interact in a hostile manner.”
The authors of this working paper consider any score between 3 and 4 to be a reasonable proxy for the concept of pernicious polarization, in which partisan identity becomes a social identity and political divisions extend into social relations. As societies divide into mutually antagonistic camps, this dynamic creates conditions in which supporters of opposing political camps are more likely to interact in a hostile rather than friendly manner.
As in any effort to quantify intangible political dynamics, this data set has limitations. It is rooted in retrospective, subjective judgments of experts and therefore is imperfect and likely incomplete. Valid criticisms have been raised against these data and, more broadly, quantitative studies of polarization. The authors recognize these concerns and address them at length in appendix C.
Nevertheless, limitations are inherent to any data set on historical levels of polarization, and tradeoffs are inherent in comparative and longitudinal research. Still, analyzing trends in polarization and depolarization dynamics within countries and comparing across countries provides the basis for identifying common patterns, exploring associated factors, and designing research studies to examine causal mechanisms. These benefits can outweigh the costs, as long as researchers are transparent about the limitations and do not make unsupported causal inferences. This study, then, represents a first attempt at understanding comparative polarization and depolarization over time, which other researchers hopefully will build upon and improve.
To classify the political systems of the countries it covers, this study also uses V-Dem’s categorization of political regimes over time through its Regimes of the World index. V-Dem created this measure with a scale of 0 to 3 as a composite of several of its indicators about electoral and liberal components of democracy (see table 1).12 Although there are many metrics of regime classification—including Freedom House’s Freedom in the World index or the Center for Systemic Peace’s Polity scale, V-Dem’s data are especially useful since they also extend to 1900, allowing for a more straightforward analysis of a country’s political regime and polarization levels in a given year. Moreover, as studies have shown, V-Dem’s data are closely aligned with other metrics in over 90 percent of cases, reflecting a high degree of agreement among them.13
|Table 1. V-Dem’s Regimes of the World Scoring System|
|Scale Rating||Regime Type||Regime Description|
|0||Closed autocracy||“No multiparty elections for the chief executive or the legislature”|
|1||Electoral autocracy||“Multiparty elections for the chief executive and the legislature, but failing to achieve that elections are free and fair”|
“Free and fair multiparty elections . . . [but deficits in] access to justice, or transparent law enforcement, or liberal principles of respect for personal liberties, rule of law, and judicial as well as legislative constraints on the executive”
|3||Liberal democracy||“Free and fair multiparty elections . . . [and] access to justice, transparent law enforcement and the liberal principles of respect for personal liberties, rule of law, and judicial as well as legislative constraints on the executive.”|
|Source: Michael Coppedge, John Gerring, Carl Henrik Knutsen, Staffan I. Lindberg, and Jan Teorell et al., “V-Dem Codebook v11.1,” Varieties of Democracy Project, 283, https://www.v-dem.net/static/website/img/refs/codebookv111.pdf.|
Global Patterns in Polarization
Though data-driven, comparative polarization research is in its early stages, a few trends stand out. First, “us versus them” polarization is increasing worldwide. Every region in the world except Oceania has seen polarization levels rise since 2005 (see figure 1). Africa has had the smallest increase during this time period, although its regional averages have been higher than those of other regions since the 1980s. Europe had a large spike during World War II and then sustained relatively low levels of polarization, until it began rising again in 2005. Latin America and Asia broadly depolarized following democratic transitions in the 1970s and 1980s, but then they began repolarizing after 2000. Oceania had the lowest regional averages until it began polarizing in the 1970s and reached similar levels to other parts of the world. The Middle East and North Africa remained at moderately high levels until polarization surged in the mid-2000s.
Breaking down the regions further reveals a more nuanced picture of the countries and subregions driving these broader trends. In Europe, for example, the Balkans, East and Central Europe, and Southern Europe are largely fueling the region’s recent rise in polarization, while Western Europe and the Nordic countries remain at comparatively low levels but have seen rises since 2005 as well (see figure 2). In South Asia, Bangladesh has been consistently polarized since its independence in 1971, while India’s polarization levels have surged since the accession of Prime Minister Narendra Modi in 2014 (see figure 3).
In the Americas, the largest democracies—Brazil, Mexico, and the United States—have been driving the recent increase in the region’s polarization levels. Figure 4 shows that the United States remained below the world and Latin American averages from 1900 until 2002; it has continued to climb above world averages ever since. As figure 5 illustrates, extremely high levels of polarization can be observed in Brazil, Mexico, and the United States in recent years—though all countries in South America except Paraguay and Uruguay are currently at pernicious levels of polarization.
Finally, what about polarization in advanced Western democracies worldwide? In this respect, the United States stands out. Figure 6 shows that, since 2005, U.S. polarization levels have skyrocketed above those of other long-standing democracies. In fact, only two other peer democracies ever came close to pernicious levels of polarization at all: France reached this benchmark for one year during the political upheaval of 1968, and Italy was just shy of pernicious levels in the 1970s during the tumultuous period known as the Years of Lead.14 Polarization in the United States, by contrast, has risen consistently since 1990 and has been at pernicious levels since 2015. The polarization of U.S. politics is more akin to the experiences of younger, less wealthy, and severely divided democracies and electoral autocracies than to those of its more consolidated democratic peers.15
There tends to be a negative (or inverse) relationship between polarization and the quality of democracy. The world experienced a global wave of democratization accompanied by depolarization following World War II, with the restoration of democracy in Europe and parts of Latin America, and independence movements in Africa and Asia. Another rise in polarization coincided with a decline in democracy in the 1960s and 1970s, followed by depolarization during the Third Wave of democratization beginning in the late 1970s. Finally, a decline in democracy can be observed after 2014 following the rise in polarization that began the decade before.
Statistical analysis confirms the existence of this relationship. A study by Murat Somer, Jennifer McCoy, and Russell Luke found that rising political polarization predicted deterioration in liberal democracy scores and that periods of sustained, severe levels of pernicious polarization have the most damaging effects on democratic quality.16 The reasons are many and may vary by country, but this general pattern of growing unwillingness to compromise and perceptions of the other side as an existential threat may encourage leaders to take steps to entrench their electoral advantage, concentrate power in the executive branch, and delegitimize critics and opponents.17 Concomitantly, depolarization is associated with improvements in democratic quality; however, periods of substantial depolarization from previously severe and pernicious levels do not add to that improvement (as pernicious polarization added to democratic erosion). This suggests that once a country reaches severe levels of polarization, this condition is uniquely bad for democracy, and simply reducing it will not necessarily help restore democratic quality.
Identifying and Classifying Depolarization Episodes
The 105 depolarization episodes covered in this working paper took place between 1900 and 2020. The full set of cases is listed in appendix A. For the purposes of this study, a depolarization episode was defined as any five-year period during which a country’s level of polarization declined from pernicious levels to below-pernicious levels (by a value of at least 0.4 on the polarization scale), without repolarizing above 3.0.18This threshold of a 0.4 decrease was chosen to ensure that all of the depolarization episodes reflected a substantial reduction of animosity between political camps.
The set of 105 depolarization episodes represents almost exactly half of the total number (211) of pernicious polarization episodes during this time period. That is to say, over 200 times between 1900 and 2020, countries reached polarization levels above 3.0 on the polarization scale for at least one year, and in half of those episodes, they were able to reduce it below pernicious levels for at least five years. Notably, the majority (58 percent) of those unresolved pernicious episodes are recent cases where countries have sustained pernicious polarization to date (like the United States), or cases in which five years have not yet elapsed to be able to assess whether they will meet the depolarization criteria or not (such as the drop in Indonesia’s polarization levels in 2019). The other 42 percent of episodes are those in which countries either remained perniciously polarized in the long run (such as Venezuela from 2002 to date or Bangladesh since independence in 1971), or more often, ones in which states repolarized within the five-year window and so do not meet the criteria for a depolarizing episode (such as Armenia, which has repolarized six times since 1996).
Another way to assess the capacity to depolarize is to look at the number of countries that have done so, rather than the quantity of episodes, because some countries have had cyclical polarization, with multiple episodes of pernicious polarization and depolarization. V-Dem presents data on political polarization for 178 countries. Of these, 117 countries (66 percent) experienced one or more episodes of pernicious polarization, but only thirty-five countries (20 percent) failed to experience any depolarizing episode from pernicious levels (again, the United States stands out as an example of this with its ongoing period of pernicious polarization beginning in 2015).
Given this apparent capacity among the majority of the world’s countries to depolarize at least some of the time from pernicious levels, the analysis here seeks to identify the contexts and sustainability of those experiences and to encourage further research into the causal mechanisms and outcomes for democracies. Thus, after identifying these depolarizing episodes, the authors sought to identify the broad political contexts of each case. To do so, the authors conducted qualitative research on each country’s political history immediately prior to and during the episode period. Based on this contextual analysis, they then determined which primary change—if any—appeared to have created enabling conditions for depolarization. They then grouped each case by these contextual factors, whether depolarization occurred after a certain type of shock or whether within a given regime type if no such shock was identified.
Some episodes could plausibly be grouped under multiple categories. India’s depolarization between 1946 and 1950, for example, took place after a violent internal conflict, regime change, and independence. However, given that the systemic political change that generated both violent conflict and a change in the country’s regime score was directly related to the country’s newfound independence, India was categorized as a case of depolarization after resolving an independence struggle; in other words, independence was the primary contextual factor that enabled depolarization.
By contrast, in cases where there was no systemic shock and a country’s regime score did not change, depolarization episodes were labeled as occurring within a given regime type.19 Future research will examine the causal mechanisms leading to depolarization in these cases. For example, a liberalizing political opening, involving a change from a more traditionally repressive leader to a more liberalizing authoritarian leader, may lead to depolarization even without a change of regime category from “closed autocracy.” Similarly, an elite political conflict may cause polarization, with depolarization occurring after one side essentially wins and becomes the dominant party, thus reducing or even repressing polarization.
Table 2 shows the breakdown of depolarizing episodes grouped by context. The five contextual categories—foreign intervention, postconflict, postindependence, post–regime change, and within a given regime type—also feature subcategories meant to add more nuance. Some of the episodes within a given regime category may also involve a significant political upheaval or change in the type of government. For example, China between 1975 and 1979 underwent dramatic political and economic changes as it transitioned from totalitarianism under Mao Zedong to a more economically and politically open form of self-described collective leadership; however, despite fundamental changes in the nature of the regime, it was coded in the same regime type since China remained a closed autocracy. Given that the total number of depolarization episodes is 105 (very close to 100), the number of cases in any context category roughly corresponds to the category’s percentage share of the total as well.
|Table 2. Contexts of Depolarizing Episodes Globally (1900–2020)|
|Category||Subcategory||Subcategory count||Total count|
|Post–regime change||Autocracy to democracy (any levels)||23||36|
|Democracy to autocracy (any levels)||2|
|Closed autocracy to electoral autocracy||9|
|Electoral autocracy to closed autocracy||2|
|Electoral democracy to or from liberal democracy||0|
|Within regime||Within closed autocracy||9||27|
|Within electoral autocracy||11|
|Within electoral democracy||7|
|Within liberal democracy||0|
|Source: Authors’ classifications based on V-Dem data.|
These data allow for several findings. First, the prevalence of systemic shocks in bringing about depolarization—whether after a civil conflict, a foreign war, regime change (primarily from autocracy to democracy), or the resolution of an independence struggle—was especially striking. Nearly three-quarters of the depolarization episodes came after such systemic disruptions. Among the systemic shocks, foreign intervention is the least common context for depolarization: there were only three such cases, and only one of them was sustained after five years. This appears to give little reason to believe that domestic political polarization can be resolved successfully and sustainably by outside powers intervening in another country’s politics.
Second, this analysis showed, perhaps not surprisingly, that countries frequently depolarize in the aftermath of violence. This most frequently came after the resolution of some sort of internal state conflict, whether a civil war, ethnic conflict, or revolutionary upheaval (one-fifth of all depolarizing episodes). Some countries, albeit a significantly smaller set, depolarized after resolving a violent conflict with another country. In such cases of internal or external conflict, the authors expect that depolarization occurs with the resolution of polarization that may have contributed to the conflict in the first place, either because one side wins and perhaps suppresses the polarization or because the two sides reach a negotiated solution.
Third, 14 percent of the cases involved depolarization after independence struggles. This is not surprising, as independence campaigns often divide society not only over the goal of independence but also over the timing and means of such dramatic political change; moreover, the resolution of such conflicts, whether by force or more peaceful political processes, can obviate these tensions. Many of these cases were drawn from the wave of colonial independence movements after World War II, though others took place in the aftermath of World War I. Violence played an important role in many of these independence-related depolarization episodes, as more than half of them came in the aftermath of armed struggles.
The remaining depolarization episodes—those that came after regime change or within a given regime type—were the most interesting. These cases are the most likely to provide policy-relevant insights for contemporary cases of pernicious polarization in democracies. A comparison of these two categories showed that depolarization periods following regime change are the most common depolarization context, constituting a plurality (34 percent) of all episodes. This finding confirmed the authors’ expectations and previous findings that pernicious polarization is a problem that may be resolved through fundamental transformations of how a society is governed and distributes power.
Analyzing the subcategories of post–regime change contexts clearly shows that autocratization is rarely followed by depolarization. There were only two cases where depolarization followed autocratization: Turkey from 1980 to 1984 and Fiji from 2012 to 2016. In sharp contrast, there were twenty-three cases (almost two-thirds of the post–regime change cases) where regime change from autocracy to democracy was followed by depolarization. Meanwhile, nine of the cases involved liberalization within autocratic regimes, a trend that points to the potential normative power of democracy. Even a relatively modest opening of democratic space—bringing about a transition from closed to electoral autocracy, but short of a transition to even electoral democracy—may have a chance of depolarizing society.
Yet not all depolarization seems to require a significant systemic shock or regime change; indeed, a quarter of the cases are episodes of depolarization within a given regime structure. This does not mean that these countries had a complete lack of political upheaval or constitutional change, just that their regime type as listed in the Regimes of the World index did not shift or else perhaps showed a short, temporary blip prior to the depolarization period before returning to its steady status in its original regime category.
Each of these depolarization cases within the same regime type exhibits its own mechanisms and strategies. Furthermore, depolarization is not always a positive occurrence. In closed autocracies, for example, depolarization may come about as political space is closed and opposition is repressed, perhaps following a contested succession struggle. In electoral autocracies, depolarization may come about after only limited reforms. In electoral democracies, by contrast, depolarization has come after political parties have brokered a deal, a contested election has been resolved, or a polarizing leader has been voted out.
Notably, there were no examples of depolarization from pernicious levels among liberal democracies. This reflects two interrelated issues. First, very few advanced democracies have suffered pernicious levels of polarization, which suggests that liberal democracy has been quite effective at preventing pernicious polarization (at least so far). At the same time, this may also suggest that it is very hard to address pernicious polarization within liberal democracy once such a country reaches that point. Instead, democratic erosion is strongly associated with pernicious polarization, and regime degradation may be a risk.19
Gauging the Sustainability of Depolarization Episodes
Having examined which contexts tend to coincide with depolarization episodes, it is worthwhile to turn to some key questions about the sustainability of depolarization. First, it is important to consider what happens in the aftermath of a depolarization episode: Is depolarization sustainable for a decade or more, or do societies tend to repolarize to pernicious levels in the short run? Second, do some contexts tend to result in more durable depolarization than others? And third, can depolarization be sustained in the long term, beyond a decade, or is polarization more commonly a cyclical pattern of politics?
Trends in the Sustainability of Depolarization From Pernicious Episodes
To determine what happened to a country’s polarization levels after the initial depolarization episode, this analysis tracked the ten years following the end of the initial depolarization period and traced the trajectory of the country’s polarization metric over that time. Four outcomes were possible (see table 3).
|Table 3. Potential Outcomes of Depolarization Episodes|
|Medium-Term Result of Depolarization||Description|
|Nonsustained||The country’s polarization level became pernicious again within ten years|
|Sustained||The country’s polarization level remained constantly below pernicious levels (3.0) and (in some cases) continued to decline over the next ten years|
|Managed||The country repolarized a substantial amount (a change of 0.4 or more on the scale) without becoming pernicious within the ten-year window or the country’s degree of polarization remained somewhat volatile at near-pernicious levels (between 2.6 and 2.99 on the scale)|
|Indeterminate||The country’s depolarization episode is too recent to measure a subsequent ten-year window|
|Source: Definitions assigned by the authors.|
A sustained depolarization episode is any period where the initial five-year decline in polarization was followed by ten years without any significant repolarization, defined as increasing by at least 0.4 on the 0 to 4 scale. For example, the Belgian episode (1945–1949) is a typical case of sustained depolarization. As figure 7 shows, after the end of World War II, Belgium’s polarization levels dropped as the country transitioned from a closed autocracy under Nazi occupation to a liberal democracy.
In managed polarization episodes, a country’s polarization level remains less than pernicious, but it either experiences substantial repolarization or remains close to pernicious—between 2.6 and 2.99 on the scale. For example, Hungary (1989–1993) is a typical case of managed polarization based on the first criterion. As figure 8 demonstrates, Hungary suffered from pernicious polarization for a very long time. Only after 1989 did the country’s polarization rate drop below pernicious, and it stayed at less-than-pernicious levels over the subsequent ten years (1994–2004). Since the country’s polarization metric increased from 1.75 in 1993 to 2.44 in 2003 (by more than 0.4), this case was categorized as managed polarization. Nevertheless, Hungary was unable to sustain this depolarization in the long run: it returned to pernicious levels of polarization in 2010 when Prime Minister Viktor Orbán’s party, Fidesz, won a supermajority in parliament and embarked on a campaign to overhaul the country’s constitution.
Turkey, on the other hand, is an example of managed polarization where the country’s polarization levels hovered at close-to-pernicious levels. As seen in figure 9, Turkey was able to depolarize after a 1980 coup, but its polarization levels remained above 2.8 until 2001. The country became perniciously polarized in 2002 when the Justice and Development Party won control of parliament.
By contrast, nonsustained depolarization occurs if a country’s polarization score returns to pernicious levels within ten years. Bolivia (2008–2012) is an illustrative example. As figure 10 shows, the country’s polarization metric remained below pernicious levels for four years but returned to pernicious levels in 2016. In 2020, the country witnessed destructive levels of polarization with an all-time high score (3.98) after a contested election in November 2019 and civilian protests forced out a controversial incumbent president, after which a divisive interim government took power and new elections were held in October 2020.20 Finally, some depolarization episodes occurred too recently to establish a subsequent ten-year trajectory, so these cases were coded as indeterminate.21
Table 4 illustrates the breakdown of the depolarization cases in terms of outcomes. While the plurality of episodes were sustained over the following decade, an equal proportion were either managed or nonsustained. This chart illustrates a recurring theme: even after falling below pernicious levels, polarization can remain a major dynamic in a country’s politics.
|Table 4. Distribution of Medium-Term Outcomes After Depolarization Episodes|
|Category||Number of episodes||Percentage|
|Sustained depolarization||48||46 %|
|Nonsustained depolarization||16||15 %|
The context in which depolarization occurs may affect how sustainable it ends up being. Table 5 presents this relationship, showing the total number of cases in each context category and the percentage of cases in that category that are sustained, managed, or nonsustained.
|Table 5. Distribution of Aftermath of Depolarization Episodes by Context|
|Category||Subcategory||Sustained (%)||Managed (%)||Nonsustained (%)||Total count|
|Post–regime change||Autocracy to democracy (any levels)||61%||30%||9%||23|
|Democracy to autocracy (any levels)||0%||100%||0%||1|
|Closed autocracy to electoral autocracy||71%||29%||0%||7|
|Electoral autocracy to closed autocracy||50%||50%||0%||2|
|Electoral democracy to or from liberal democracy||N/A||N/A||N/A||0|
|Within regime||Within closed autocracy||33%||33%||33%||9|
|Within electoral autocracy||20%||50%||30%||10|
|Within electoral democracy||40%||40%||20%||5|
|Within liberal democracy||N/A||N/A||N/A||0|
|Note: This table omits the indeterminant category of seven cases too recent for assessing sustainability, so the total cases reported here is 98. (These seven cases are listed in table 8 at the end of appendix C.)|
The data set does not allow the authors to draw causal inferences, but the findings suggest directions for future research. The depolarization contexts that appear to yield the most sustainable change are those where politics becomes more open, whether after a democratic transition or even liberalization within an authoritarian regime. The resolution of violent intrastate conflict also frequently results in sustained depolarization. These findings offer tentative evidence that competitive politics—or democracy, depending on how the findings are interpreted—and an end to widespread violence are two effective ways of addressing polarization.
Long-Term Outcomes of Depolarization Episodes
Although it is encouraging to see that the majority of cases show sustained depolarization or managed polarization over the first ten-year period that followed, it is worth asking: How many of these cases were able to avoid pernicious polarization beyond that initial ten-year window? Each episode was examined until 2020 (the last year available in the V-Dem data set when this analysis was conducted). The findings offer a somewhat hopeful picture. Out of forty-eight episodes where depolarization was sustained for at least a decade, twenty-nine cases (about 60 percent) have avoided repolarizing to pernicious levels to date. Even more encouraging, table 6 shows that half of these long-term depolarizers have been able to keep their polarization levels below 2.0, well below the pernicious threshold of 3.0.
|Table 6. Sustained, Long-Term Depolarization (up to 2020)|
|Country||Depolarization Episode Years||Below 2.0 Until 2020||Between 2.0 - 3.0 Until 2020|
|Source: See Coppedge, Gerring, Knutsen, Lindberg, and Teorell et al., “V-Dem Dataset Version 11.1,” V-Dem, March 2021, https://www.v-dem.net/en/data/archive/previous-data/v-dem-dataset.|
In contrast, out of the thirty-four cases of managed polarization, fewer than half (41 percent) were able to avoid repolarizing to pernicious levels, demonstrating their vulnerability to repolarization (see table 7). Together, nearly half (thirty-nine of eighty-two) of the sustained and managed episodes repolarized to pernicious levels in the long run, suggesting that there may be a strong cyclical element to pernicious polarization.
|Table 7. Managed, Long-Term Depolarization|
|Country||Depolarization Episode Years||Below 2.0 Until 2020||Between 2.0 and 3.0 Until 2020|
|Source: Compiled by the authors’ based on V-Dem data.|
The cyclical nature of polarization is made even clearer by another group: countries that have experienced pernicious polarization multiple times after their initial depolarization episode. This vicious cycle of polarization, depolarization, and repolarization has occurred frequently in seventeen countries. Bolivia, Burundi, Colombia, the Democratic Republic of Congo, France, Hungary, Iran, Peru, Serbia, South Africa, Timor-Leste, Turkey, Uruguay, and Venezuela became perniciously polarized at least two more times; Cambodia and Russia repolarized to pernicious levels three more times; and Senegal did so a whopping four more times.
This working paper represents a first attempt to identify and analyze a collection of depolarization cases. Although the data set did not support causal inferences, this study established a few key findings.
The analysis demonstrates that countries can successfully depolarize from pernicious levels of polarization and remain depolarized for at least five years, as illustrated by the 105 such episodes globally since 1900. The vast majority (78 percent) of these experiences were sustained or managed for a subsequent decade. When looking at the long term, however, the picture becomes less rosy: nearly half (thirty-nine) of the eighty-two sustained and managed episodes repolarized to pernicious levels in the long run. The managed cases of polarization showed the most vulnerability to repolarizing to pernicious levels in the long run.
Only a fraction (14 percent) of the depolarizing episodes resulted in sustained low levels (below 2.0) of polarization in the long term (at least to date). Investigating the mechanisms and strategies that enable such sustained depolarization within a democratic context will be the subject of future research. But given the small number of democracies (eleven) able to accomplish this feat amid the larger pattern of cyclical polarization and depolarization, it also will be crucial to understand strategies of managing polarization at moderately high levels while avoiding democratic erosion, government dysfunction, or returns to pernicious polarization and potential violence.
This analysis also points to the strong role of democratizing experiences in depolarizing societies. The global waves of democratization accompanying the defeats of fascism and communism, decolonization, and the end of military authoritarian dictatorships in the twentieth century provided the context for many of the depolarizing episodes. On the other hand, only seven episodes of depolarization occurred within an electoral democracy (and none within a liberal democracy).
History therefore offers fewer cases to learn from for addressing a major challenge the world faces in the twenty-first century: democracies are increasingly suffering from pernicious polarization. The United States poses a particularly troubling case, as it is the only advanced Western democracy that has suffered such high levels of polarization for such an extended period. This experience, along with the recent high levels of polarization in other large democracies, such as Brazil, India, and Turkey, points to the urgent need not only to learn from the past but also to innovate new mechanisms to reduce or better manage this phenomenon.
Appendix A: List of All Depolarization Episodes (1900–2020)
The table below shows the full list of 105 depolarization cases that appear in this working paper.
|Country||Depolarization episode years|
|Democratic Republic of the Congo||1965-1969|
|Democratic Republic of the Congo||2002-2006|
Appendix B: List of All Depolarization Episodes by Context Group
The table below categorizes all the depolarization episodes by context.
|Category||Subcategory||Number of cases|
|Democratic Republic of the Congo||1965-1969||Managed|
|Democratic Republic of the Congo||2002-2006||Nonsustained|
|Post–regime change||Autocracy to democracy (any levels)||23|
|Post–regime change||Democracy to autocracy (any levels)||2|
|Post–regime change||Closed autocracy to electoral autocracy (any levels)||9|
|Post–regime change||Electoral autocracy to closed autocracy (any levels)||2|
|Within regime||Closed autocracy||9|
|Within regime||Electoral autocracy||11|
|Within regime||Electoral democracy||7|
Appendix C: Features and Limitations of the Data
V-Dem’s methodology and data set must be critically evaluated to establish the data’s explanatory power. This appendix aims to describe V-Dem’s research methodology and identify some of its limitations.
V-Dem’s data set is based on a decentralized research design in which country experts are surveyed online to quantitatively assess a given country’s political situation.23 Country experts offer a rich data source, but they also introduce problems of subjectivity. These may be related to any number of factors, including cultural and professional biases, personal experiences, or ideological outlooks. Moreover, given that both regime type and polarization are latent concepts (meaning that they are not directly observed but inferred), concerns associated with country expert surveys may increase.
V-Dem has adopted a method using Bayesian item response theory to overcome challenges of intercoder reliability for a single country. This approach helps scholars to convert raw data coded in ordinal measure (these are variables that have an order or rank but without a degree of difference between categories) to point estimates for each indicator, country-year observation, and their associated measures of uncertainty (confidence intervals and standard deviations). Thus, Bayesian item response theory allows V-Dem to correct for any bias that might stem from country experts through this conversion and inclusion of uncertainty measures.
Yet this approach cannot solve issues stemming from how the expert coders perceive latent variables across countries. While the option of having a single set of in-house expert coders to assess all countries, or even a single region, could relieve some of the concerns of variation across groups of experts, such an approach would sacrifice some of the deep expertise of country experts. V-Dem tries to partially resolve this by having bridge coders who have cross-country and intertemporal expertise. In bridge coding, a single expert is instructed to code multiple countries for all years. In comparison to a coder who rates multiple countries for a limited period, bridge coders can explicitly compare different political contexts and periods. Additionally, this approach helps the bridge coder measure the variable correctly by avoiding potential systematic biases that might result from employing different standards in measuring. There are 700 bridge coders (25 percent of all experts), and these experts code 2.5 countries on average.24
Another issue stems from the retrospective evaluation of political dynamics. Country experts are likely to suffer from hindsight bias, “the tendency to retrospectively exaggerate one’s foresight of a particular event.”25 This tendency can be more problematic if an expert coder cannot differentiate their current knowledge that an event occurred from their assessment of the period before that event.
V-Dem is aware of these limitations: to mitigate the outsized effects of individual ratings in small samples, it recommends not using score estimates for individual countries in individual years with three or fewer expert assessments.26 In the identified depolarization episodes, there are five episodes with three or fewer experts: Brazil (1989–1993), Finland (1918–1922), Mali (1990–1994), Timor-Leste (1999–2003), and Timor-Leste (2007–2011). These cases were kept in the data set because the study’s qualitative research also suggested these were cases of depolarization.
The only case in the study’s qualitative desk research on each country’s political history where major contextual factors that may have contributed to the depolarizing dynamic could not be identified or corroborated is Cyprus (1989–1993). While these years also coincide with the Cyprus conflict, which has been an ongoing dispute since the 1960s, the political trend of these years could not be attributed to any particular development in this conflict. Thus, this episode’s context is coded as electoral democracy, and it requires further investigation and input from country experts.
In addition to these categories, there were a number of cases that could not be evaluated due to an insufficient time horizon. It is too early to categorize depolarization episodes that started in or after 2007 as to whether they were able to sustain depolarization or manage polarization for another decade. In total, there were seven recent cases that could not be categorized (see table 8).
|Table 8. Indeterminate Recent Cases|
|Country||Depolarization episode years||Depolarization range|
|Burma||2007-2014||3.82 to 1.84|
|Egypt||2014-2019||3.84 to 2.86|
|Ethiopia||2007-2020||3.47 to 2.61|
|Fiji||2002-2018||3.80 to 2.35|
|Mongolia||2008-2016||3.22 to 2.06|
|Senegal||2012-2016||3.19 to 1.64|
|Timor-Leste||2007-2016||3.05 to 1.50|
The Carnegie Endowment for International Peace is grateful to the Ford Foundation, whose generous support helped make the research for this paper possible. The authors appreciate the helpful comments from members of Carnegie’s Democracy, Conflict, and Governance Program as well as Dan Slater and Adrienne LeBas on an earlier version of this paper.
1 Thomas Carothers and Andrew O’Donohue, eds., Democracies Divided: The Global Challenge of Political Polarization (Washington, DC: Brookings Institution Press, 2019); Stephan Haggard and Robert Kaufman, Backsliding: Democratic Regress in the Contemporary World (Cambridge: Cambridge University Press, 2021); Steven Levitsky and Daniel Ziblatt, How Democracies Die (New York: Crown, 2018); and Jennifer McCoy and Murat Somer, eds., “Special Issue on Polarized Polities: A Global Threat to Democracy,” Annals of the American Academy of Political and Social Science 681, no. 1 (January 2019) https://www.aapss.org/volumes/polarizing-polities-a-global-threat-to-democracy/.
2 Nancy Bermeo, Ordinary People in Extraordinary Times: The Citizenry and the Breakdown of Democracy (Princeton, NJ: Princeton University Press, 2003); Adrienne LeBas, From Protest to Parties: Party-Building and Democratization in Africa (New York: Oxford University Press, 2011); Jennifer McCoy, Tahmina Rahman, and Murat Somer, “Polarization and the Global Crisis of Democracy: Common Patterns, Dynamics, and Pernicious Consequences for Democratic Polities,” American Behavioral Scientist 62, no. 1 (January 1, 2018): 16–42; and Kenneth M. Roberts, “Populism and Polarization in Comparative Perspective: Constitutive, Spatial and Institutional Dimensions,” Government and Opposition (June 7, 2021): 1–23.
3 McCoy, Rahman, and Somer, “Polarization and the Global Crisis of Democracy.”
4 Donald P. Green, Bradley Palmquist, and Eric Schickler, Partisan Hearts and Minds: Political Parties and the Social Identities of Voters, Yale ISPS Series (New Haven: Yale University Press, 2002); Lilliana Mason, “‘I Disrespectfully Agree’: The Differential Effects of Partisan Sorting on Social and Issue Polarization,” American Journal of Political Science 59, no. 1 (January 2015): 128–145; and Henri Tajfel and J. C. Turner, “An Integrative Theory of Intergroup Conflict,” in The Social Psychology of Intergroup Relations, edited by W. G. Austin and S. Worchel (Monterey, CA: Brooks/Cole, 1979), 33–37.
5 Mason, “‘I Disrespectfully Agree.’”
6 Murat Somer, Jennifer McCoy, and Russell Luke, “Pernicious Polarization, Autocratization and Opposition Strategies,” Democratization 28, no. 5 (2021): 929–948; and Jennifer McCoy and Murat Somer, “Pernicious Polarization and Democratic Resilience in the US: Comparative Lessons,” in Democratic Resilience: Can the United States Withstand Rising Polarization? edited by Robert C. Lieberman, Suzanne Mettler and Kenneth Roberts, (Cambridge: Cambridge University Press, 2021).
7 Murat Somer and Jennifer McCoy, “Déjà vu? Polarization and Endangered Democracies in the 21st Century,” American Behavioral Scientist 62, no. 1 (January 1, 2018): 3–15.
8 Murat Somer and Jennifer McCoy, “Transformations Through Polarizations and Global Threats to Democracy,” Annals of the American Academy of Political and Social Science 681, no. 1 (January 2019): 8–22, https://doi.org/10.1177/0002716218818058.
9 Jennifer McCoy and Murat Somer, eds., “Special Issue: Polarization and Democracy: A Janus-Faced Relationship With Pernicious Consequences,” American Behavioral Scientist 62, no.1 (2018) https://journals.sagepub.com/toc/abs/62/1; and McCoy and Somer, eds., “Special Issue on Polarized Polities: A Global Threat to Democracy”; Carothers and O’Donohue, Democracies Divided; and Haggard and Kaufman, Backsliding.
10 The Varieties of Democracy (V-Dem) Institute is an institution hosted by the Department of Political Science at the University of Gothenburg (in Sweden). While V-Dem’s efforts to gather an extensive data set on democracy date back to 2011, the full V-Dem data set was released in 2016. This research uses version 11.1 of the data set, which was released in 2021. See Michael Coppedge, John Gerring, Carl Henrik Knutsen, Staffan I. Lindberg, and Jan Teorell et al., “V-Dem Dataset Version 11.1,” V-Dem, March 2021, https://www.v-dem.net/en/data/archive/previous-data/v-dem-dataset/.
11 Michael Coppedge, John Gerring, Carl Henrik Knutsen, Staffan I. Lindberg, and Jan Teorell et al., “V-Dem Codebook (v11.1),” Varieties of Democracy (V-Dem) Project, March 2021, https://www.v-dem.net/static/website/img/refs/codebookv111.pdf.
12 For more on the Regimes of the World index, see Coppedge, Gerring, Knutsen, Lindberg, and Teorell et al., “V-Dem Codebook (v11.1),” 283.
13 Existing regime classifications cannot be used interchangeably, and the research outcomes are highly dependent on the selection of measures. Nevertheless, the level of agreement on categorizing regimes between V-Dem and other data sets is more than 90 percent. For further discussion, see Gretchen Casper and Claudiu Tufis, “Correlation Versus Interchangeability: The Limited Robustness of Empirical Findings on Democracy Using Highly Correlated Datasets,” Political Analysis 11, no. 2 (2003): 196–203; John Högström, “Does the Choice of Democracy Measure Matter? Comparisons Between the Two Leading Democracy Indices, Freedom House and Polity IV,” Government and Opposition 48, no. 2 (2013): 201–221; and Espen Geelmuyden Rød, Carl Henrik Knutsen, and Håvard Hegre, “The Determinants of Democracy: A Sensitivity Analysis,” Public Choice 185, no. 1 (2020): 87–111. A comparison of RoW with regime indices developed by Boix et al. (2013), Cheibub et al. (2010), Geddes et al. (2014), Wahman et al. (2013), the Polity score, and Freedom House shows that the RoW index is more conservative if there is a disagreement between them, since RoW’s categorization of democracy has a higher bar compared to others. For comparison of these measurements, see Anna Lührmann, Marcus Tannenberg, and Staffan I. Lindberg, “Regimes of the World (RoW): Opening New Avenues for the Comparative Study of Political Regimes,” Politics and Governance 6, no.1 (2018): 60. For these alternative measurements, see Carles Boix, Michael Miller, and Sebastian Rosato, “A Complete Data Set of Political Regimes, 1800–2007,” Comparative Political Studies 46, no. 12 (2013): 1523–1554; José Antonio Cheibub, Jennifer Gandhi, and James Raymond Vreeland, “Democracy and Dictatorship Revisited,” Public Choice 143, no. 1 (2010): 67–101; Barbara Geddes, Joseph Wright, and Erica Frantz, “Autocratic Breakdown and Regime Transitions: A New Data Set,” Perspectives on Politics 12, no. 2 (2014): 313–331; Michael Wahman, Jan Teorell, and Axel Hadenius, “Authoritarian Regime Types Revisited: Updated Data in Comparative Perspective,” Contemporary Politics 19, no. 1 (2013): 19–34; Monty G. Marshall, Keith Jaggers, and Ted R. Gurr, “Polity IV Individual Country Regime Trends, 1946–2013,” Center for Systemic Peace, 2014, https://www.systemicpeace.org/polity/polity4.htm; and Freedom House, “Freedom in the World 2016,” Freedom House, https://freedomhouse.org/sites/default/files/FH_FITW_Report_2016.pdf.
14 “Italian Neofascism and the Years of Lead: A Closer Look at the Nuclei Armati Rivoluzionari,” Middlebury Institute of International Studies at Monterey, March 4, 2022, https://www.middlebury.edu/institute/academics/centers-initiatives/ctec/ctec-publications/italian-neofascism-and-years-lead-closer-look.
15 Jennifer McCoy and Benjamin Press, “What Happens When Democracies Become Perniciously Polarized?” Carnegie Endowment for International Peace, January 18, 2022, https://carnegieendowment.org/2022/01/18/what-happens-when-democracies-become-perniciously-polarized-pub-86190.
16 Somer, McCoy, and Luke, “Pernicious Polarization, Autocratization and Opposition Strategies.”
17 Haggard and Kaufman, Backsliding; and Jennifer McCoy and Murat Somer, “Toward a Theory of Pernicious Polarization and How It Harms Democracies,” Annals of the American Academy of Political and Social Science 681, no. 1 (December 20, 2018): 234–271, https://journals.sagepub.com/doi/full/10.1177/0002716218818782.
18 Each five-year window started in the last year during which a country was perniciously polarized.
19 McCoy and Press, “What Happens When Democracies Become Perniciously Polarized?”
20 Tom Phillips, “Bolivia’s New Leftwing President: ‘We Have Reclaimed Democracy,’” Guardian, November 8, 2020, https://www.theguardian.com/world/2020/nov/08/luis-arce-bolivia-president-elect-left-morales.
21 A detailed explanation of the indeterminate category can be found at the end of appendix C.
22 Coppedge, Gerring, Knutsen, Lindberg, and Teorell et al., “V-Dem Codebook (v11.1).”
23 Daniel Pemstein, Eitan Tzelgov, and Yi-ting Wang, “Evaluating and Improving Item Response Theory Models for Cross-National Expert Surveys,” V-Dem Working Paper no. 1, 2015.
24 Michael Coppedge, John Gerring, Carl Henrik Knutsen, Staffan I. Lindberg, and Jan Teorell et al., “V-Dem Methodology (v11.1),” March 2021, https://www.v-dem.net/static/website/img/refs/methodologyv111.pdf.
25 Laura Levick and Mauricio Olavarria-Gambi, “Hindsight Bias in Expert Surveys: How Democratic Crises Influence Retrospective Evaluations,” Politics 40, no. 4 (2020): 494–509.
26 Coppedge, Gerring, Knutsen, Lindberg, and Teorell et al., “V-Dem Codebook (v11.1).”