How democratically elected politicians get away with autocratizing their country

IDENTITY, PARTISANSHIP, POLARIZATION How democratically elected politicians get away with autocratizing their country

IDENTITY, PARTISANSHIP, POLARIZATION Authors Elena Avramovska is a senior researcher for democracy, society and youth at the Friedrich-Ebert-Stiftung’s(FES) Office ‘Democracy of the Future’. Her research has been funded by the European Commission, the Swiss National Research Foundation and the Open Society Foundation. Avramovska has contributed to publications with Oxford University Press and Routledge. Johanna Lutz is the director of the Friedrich-Ebert-Stiftung’s(FES) Office‘Democracy of the Future’. She has worked on democratization and democracy support also in her previous roles as Deputy Head of the FES EU Office in Brussels and as President of the European Network of Political Foundations in Brussels. Filip Milačić is a senior researcher for democracy and society at the Friedrich-Ebert-Stiftung’s(FES) Office‘Democracy of the Future’ and a research affiliate at the Central European University’s(CEU) Democracy Institute. His new book,‘Stateness and Democratic Consolidation: Lessons from Former Yugoslavia’, was published by Springer in 2022. Milan Svolik is Professor of Political Science at Yale University. He studies the politics of authoritarian regimes, democratization and democratic backsliding. Svolik’s research has won awards and funding from the American Political Science Association, the Guggenheim Foundation and the National Science Foundation. Acknowledgments The authors’ gratitude goes first and foremost to the other members of the Democracy of the Future team, Krisztina Stefán and Michael Jennewein, for their valuable help and assistance. Furthermore, we owe debts of gratitude to the members of our advisory board, Licia Cianetti, Joanna Fomina and Uwe Optenhögel, for their valuable comments and remarks. This report also owes to Florian Müller, not only for providing us with an excellent layout and outstanding graphics for this study, but also for showing a lot of patience with us. We thank Andreas Schedler for organizing the presentation of our early findings at the Democracy Institute of the Central European University(CEU), as well as the participants of the event. We also thank Simon Richter and Austin Jang for their valuable statistical analysis. 3

How democratically elected politicians get away with autocratizing their country Table of Contents Foreword 4 Executive Summary 6 Introduction 8 Research Design 10 Comparative Part 16 Conclusion 34 Country Profiles 37 Estonia 38 Germany 44 Poland 50 Serbia 56 Spain 62 Sweden 68 Ukraine 74 References 81 Imprint 84 4

IDENTITY, PARTISANSHIP, POLARIZATION Foreword By Martin Schulz President of the Friedrich-EbertStiftung Former President of the European Parliament “Increasing polarization threatens democracy.” – Recently, this statement has become a widespread opinion that you hear and read in the media all the time, and to some degree, it seems to be backed up by scientific research. But how exactly does“polarization” threaten democracy? What are the mechanisms behind it that endanger the functioning of democracies? And does this assumption really hold true universally? Does polarization in the US work in similar ways as in Brazil, India or Poland? We do indeed observe more and more that democracies are creepingly undermined by ruling politicians. While the population has the opportunity to sanction this behavior at the next election at the latest, it does not always do so. Milan Svolik(Yale University) has used an election experiment for the United States to show that a strong polarization of society leads to a situation in which citizens no longer act as a corrective, but instead rather focus on their own camp’s victory – even if they have to vote for a politician who explicitly announces that they want to lower certain democratic or constitutional standards. Our team from the FES Office‘Democracy of the Future’ cooperated with Svolik to analyze for seven selected European countries under which circumstances the electorate acts as a corrective against the erosion of democracy. The results of our large-scale experiment in Germany, Estonia, Poland, Serbia, Spain, Sweden and Ukraine are extremely exciting. Not only do they allow us to draw conclusions about which factors may be decisive for autocratization in Europe, but we can also designate which democratic standards people are most willing to give up and for which concrete interests they are most willing to accept a curtailment of democratic standards. All in all, we can see that – despite different levels of democratic development – the mechanisms that underlie the progression of the autocratization process do work similarly in all of the selected countries. How polarized a certain society is also plays a role in our selected European countries, but not as clearly as was the case in the US experiment. This is due to the fact that the US has a two-party system which collapses a wide range of social and political debates into a singular, two-poled battle. There, a stance on a particular issue can always be directly assigned to one of the two camps, which is also reflected on the ballot in the form of a party. In turn, in European multiparty systems, polarization alone cannot have as strong an effect as it does in the United States. 5

How democratically elected politicians get away with autocratizing their country This insight is important when it comes to developing strategies to make democracies more resilient. As in the US, however, loyalty to a party plays a decisive role in Europe, too. But some issues, mostly identity-related(such as the rights of same-sex couples or immigration) ones, are sometimes considered so important that voters accept that democratic standards are dismantled bit by bit. These issues are not always polarizing, even in Europe. Moreover, an interesting finding lies in the triad of identity, polarization and party loyalty: Parties play a significant role in protecting democracy. This insight seems not new – but we were able to qualitatively and quantitatively underpin it by our findings and our figures. Our data show that the danger to democracies in Europe comes mostly from right-wing parties. But voters of some center-right parties also show that they not only tolerate the dismantling of democracy, but even approve it. It would be even more interesting to investigate why exactly party loyalty is more significant for people than the preferred policy position of a candidate. We know from previous research that group affiliation plays a very important role for political behavior and that, at the same time, parties are the most important political groups in democracies. The behavior of the“voters” in our experiment suggests that parties are attributed a certain stabilizing and corrective function that voters rely on. Thus, if a hypothetical candidate of a social democratic party proposes a clearly undemocratic measure, voters rely on the party to correct such an outlier and stick with the candidate because of other proposed policies that are important to them. Hence, an important take-away from this study is the urgent appeal to parties to become aware of their protective and controlling function for our democracies and to act accordingly. 6

IDENTITY, PARTISANSHIP, POLARIZATION EXECUTIVE SUMMARY This study examines what motivates voters to support candidates who violate democratic norms and how undemocratic politicians get away with autocratizing their countries. Identitydemocracy trade-off Voters show a greater willingness to disregard democratic violations if their preferred identitybased policies are proposed. What motivates voters to support candidates who violate democratic norms? This question is particularly pertinent in light of the ongoing trend of autocratization across Europe. In recent years, incumbents have gradually undermined democratic institutions and shifted power relations to their advantage. Despite this, many such politicians continue to enjoy the support of their constituents. This study examines what motivates voters to support candidates who violate democratic norms and how undemocratic politi cians get away with autocratizing their countries. Based on a commissioned representative survey with 10,001 interviewees in Germany, Poland, Sweden, Spain, Estonia, Ukraine and Serbia, which included a questionnaire query and candidate choice experiments, we are able to draw the following conclusions: • Most voters understand democracy Our findings reveal that respondents in the seven countries under investigation have a solid understanding of what democracy is and what it is not. • Voters are not indifferent to violations of democratic norms The majority of voters cherish democracy and punish candidates who violate democratic norms. The respondents, however, do not value all elements of liberal democracy equally. Candidates who incite violence to undermine electoral fairness(such as violently disrupting rallies) are most likely to be punished at the ballot box. Violations of checks and balances(e.g., passing laws without parliamentary debate) and civil liberties(e.g., banning foreign funding for NGOs and labor unions) receive less punishment. • Voters are partisan We find that people are partisan above everything else. Most voters forgive politicians who violate democratic norms if they represent the voters’ preferred party. This serves as proof of the ongoing strength of political parties, but also as a reminder of how important it is for party organizations to keep undemocratic politicians in check. It is also a call on all voters(partisan or non-partisan) to act as watchdogs for democracy. • Voters engage in identity-democracy trade-offs Citizens may neglect democratic principles in favor of policy preferences and partisan allegiances. Specifically, voters tend to trade off democratic values for their preferred identity policies(e.g., same-sex couples’ rights and immigration). When it comes to these issues, voters do rather disregard democratic violations than have their“camp” lose. The purely socio-economic interests do not seem to have that same strong influence on voters. According to our study, it appears that representing the voters’ preferred identity policies protects politicians from losing votes, even when they violate democratic principles. • Identity-democracy trade-offs are heightened by polarization In addition to uncovering that voters tend to 7

How democratically elected politicians get away with autocratizing their country trade off democratic norms for their preferred identity policies, our study also finds compelling evidence that identity-related issues are polarizing societies significantly more than socio-economic disagreements(e.g., on taxes). • Same-sex couples’ rights is the most polarized issue Albeit often not the most important issue to voters, same-sex couples’ rights is the most polarizing issue on average in our case countries. In combination with a high willingness to forgive undemocratic behavior by both rightand left-leaning voters, this suggests the potential of this issue to undermine democracy. • The far-right is a danger to democracy Voters of far-right parties are the least willing to punish undemocratic conduct and are the most partisan(most reluctant to punish their favorite party’s candidates when they violate democratic principles). As a result, our survey finds no evidence to support the claim that the far-right and the far-left pose equal threats to democracy. We can see that right-wing or far-right party supporters are far more conducive to undermining democracy than party supporters of left-wing or far-left parties. • Left-leaning voters play an ambivalent role While on average, voters with left-leaning positions regarding social and economic issues are far more willing to punish undemocratic conduct in general and across most countries, these voters nevertheless do play an ambivalent role: As soon as their preferred policies on identity issues are at play(e.g., the rights of same-sex couples) socially left-leaning voters show a strong tendency to trade off democratic principles for their preferred policy. • Polarization is a driver for autocratization – but not the only one In contrast to the situation in the US, voters in the countries studied were not the most likely to forgive a candidate’s undemocratic positions on issues that were intensely polarizing. The salience of an issue can also make people tolerate undemocratic behavior even if a society is not polarized over it. • Identity issues could become gateways for democratic erosion in Western Europe It seems that identity policies are a consistent determinant for a heightened willingness to accept democratic violations, and our study found that this effect is not regionally contained either. The trade-off between democracy and identity policies by which politicians are potentially able to autocratize their countries holds true everywhere, from Ukraine to Sweden. This finding opposes culturally deterministic assumptions on democratic backsliding and indicates instead that some politicians in the Eastern European countries have better understood and made more efficient use of these mechanisms for their own interests. • Voters’ tendency to punish undemocratic politicians correlates with higher levels of education and a greater understanding of democracy We found that respondents with higher levels of education, higher levels of interest in politics, an inherent opposition to strong leaders and a greater understanding of democracy punish undemocratic politicians more than other social groups. In light of this, it is necessary to encourage civic education in relation to democracy as well as its various elements. Party loyalty trumps democracy Most voters forgive politicians who violate democratic norms if the politician represents the voters’ preferred party. 8

IDENTITY, PARTISANSHIP, POLARIZATION INTRODUCTION Why do voters support candidates who violate democratic norms? Democratic backsliding Despite a solid commitment of citizens to democracy, many democracies have been backsliding. Even though in many surveys across Europe and elsewhere, large majorities of the respective populations showed a solid commitment to democracy, many countries have been experiencing democratic backsliding. Indeed, the recent erosion of democratic institutions followed a gradual process, with incumbents slowly undermining democratic institutions and tilting the balance of power to their advantage. Politicians such as Hungary’s prime minister Viktor Orbán, India’s prime minister Narendra Modi and Serbia’s president Aleksandar Vučić, for instance, have repeatedly violated the principles of liberal democracy, yet, they continue to enjoy the support of their constituencies. Hence, in our survey and research, we want to find out the following: What motivates voters to support political candidates who violate democratic norms? What circumstances support the continued support of politicians from their electorate and how can they get away with autocratizing their countries? It has already been recognized in the literature(Diamond 1994; Weingast 1997) that citizens are essential in preventing democratic backsliding. Indeed, a resilient democracy needs a critically thinking public that acts as a watchdog against violations of democratic principles. Whenever parliaments pass an undemocratic law or whenever a government enacts rules in an undemocratic and/or discriminatory way, citizens have various options to express their disagreement, for example by organizing mass protests or by starting an online petition against the action. Eventually they have the option to vote the politicians out of government in the next election. The recent waves of democratic erosion, however, show that this ideal process is not always working. Therefore, we investigate the circumstances under which citizens cease to serve as watchdogs of democracy and forgive or even reward politicians for violating democratic norms. In particular, this research aims at investigating if citizens have a good understanding of democracy and whether they value democratic principles enough to punish politicians who violate them. The study is based on representative surveys of seven European countries: Estonia, Germany, Poland, Serbia, Spain, Sweden and Ukraine. Each survey was composed of two parts: direct questions and a candidate choice experiment. 9

How democratically elected politicians get away with autocratizing their country STRUCTURE OF THE STUDY The structure of our study is as follows: In the Comparative Part, we provide a comparative analysis of the results in all seven countries examined. Results presented in this section focus on the following aspects: 1) Citizens’ support for democracy. Here, we examine whether polled citizens value democracy and have a good understanding of it; 2) We examine whether voters would punish politicians who violate democratic principles. Furthermore, we delve deeper to understand for which dimensions of liberal democracy citizens would punish undemocratic candidates the most: civil liberties, electoral fairness or checks and balances; 3) We analyze for which policy preferences or partisan interests our respondents are most willing to overlook democratic violations and to trade-off democratic principles. Morover, we identify the share of voters who are partisans first and democrats second, or, in other words, who show a strong willingness to trade off democratic principles for partisan interests and priorities. 4) We analyze if the willingness to forgive politicians who violate democratic principles is further heightened by polarization. In this context, we identify concrete polarizing issues that menace democracy, and we analyze their nature(e.g., identity-based issues vs. socio-economic issues); 5) Lastly, we offer closer insights into characteristics of our respondents, such as party affiliation, the demographic they belong to, and socioeconomic characteristics. Country Profiles: The second part of the study offers detailed analyses of each country individually: Estonia, Germany, Poland, Serbia, Spain, Sweden and Ukraine. 10

IDENTITY, PARTISANSHIP, POLARIZATION RESEARCH DESIGN CASE SELECTION 1 State of Democracy to reflect different levels of democratic development 2 Level of Polarization according to the Varieties of Democracy data, in all selected countries one can detect an increase or a high degree of political and societal polarization in the last 10 years 3 Geographical Balance to ensure the representation of the whole European continent (excluding fully authoritarian countries) The study focuses on seven European cases: Estonia, Germany, Poland, Serbia, Spain, Sweden and Ukraine. Our selection criteria for these countries were threefold: the level of democracy, the degree of polarization, and the geographical location. To examine whether there is a relationship between a country’s state of democracy and its citizens’ willingness to trade off democratic principles for other political objectives, we looked at countries with different levels of democratic development. Moreover, in order to examine whether polarization contributes to the support for undemocratic politicians, we selected countries where polarization has been increasing over the past ten years(according to the Varieties of Democracy societal and political polarization index). Finally, we selected cases from different regions of Europe. With these criteria in mind, we chose Germany as a representative of the‘old’, consolidated/ liberal democracies in Western Europe; Sweden as a representative of the‘old’, consolidated/liberal democracies in Northern Europe; Spain as a representative of the South European consolidated/liberal democracies that have been experiencing socio-economic instability during the period under observation; Estonia as a representative of the‘newly’ consolidated/liberal Baltic democracies; Poland as a representative of the Central European democracies(formerly consolidated/liberal, but characterized by democratic backsliding within the period under observation); Serbia as a representative of the Southeast European non-consolidated democracies and, finally, Ukraine as a representative of the East European non-consolidated democracies. DATA COLLECTION Our data set is based on a commissioned representative survey composed of two parts: a questionnaire query and a candidate choice experiment. The survey was conducted by Ipsos between October 21, 2021 and November 17, 2021. While for Estonia, Germany, Poland, Spain and Sweden the population under investigation was between 18 and 75 years of age, for Serbia and Ukraine the age range was 18 to 65. Interviews were conducted online and had an average length of 25 minutes. Quota selection from the Ipsos Access Panels was done according to the following criteria of representativeness: gender, age, region and education(best effort) 1 . In total, 10,001 interviews were conducted: 2,500 interviews for Germany; 1,500 interviews for Poland, Spain, and Ukraine; 1,000 interviews for Estonia and Sweden; and 1,001 interviews for Serbia. We chose online interviews because the visual options allowed us to best implement the candidate choice experiment. We are aware that for Ukraine the results of the survey might differ significantly if conducted today, with the newest experience of the ongoing Russian war against the country. Nevertheless, with regard to the recent events of Ukraine becoming a candidate country for EU membership, our data can give valuable insight into the understanding of democracy and the rule of law in Ukraine prior to the Russian invasion in February 2022. 1 In the case of Serbia and Ukraine it was not possible to meet the education criterion. Moreover, in both countries, the age category“between 60 and 65 years old” is strongly underrepresented. 11

How democratically elected politicians get away with autocratizing their country SELECTED COUNTRIES 1 Estonia as a representative of ‘newly’ consolidated/liberal Baltic democracies 2 Germany as a representative of ‘old’, consolidated/liberal democracies in Western Europe 3 Poland as a representative of Central European democracies, formerly consolidated/ liberal but experiencing democratic backsliding within the period under observation 6 1 3 2 7 4 5 4 Serbia as a representative of non-consolidated democracies in Southeastern Europe 5 Spain as a representative of South European consolidated/liberal democracies experiencing socioeconomic instability within the period under observation 6 Sweden as a representative of ‘old’, consolidated/liberal democracies in Northern Europe 7 Ukraine as a representative of non-consolidated democracies in Eastern Europe Age Sample (10,001 total) Estonia 18- 75 1,000 Germany 18- 75 2,500 Poland 18- 75 1,500 Serbia 18- 65 1,001 Spain 18- 75 1,500 Sweden 18- 75 1,000 Ukraine 18- 65 1,500 12

IDENTITY, PARTISANSHIP, POLARIZATION CANDIDATE CHOICE EXPERIMENTS Besides standard survey questions on the support for democracy, policy and party preferences, interest in politics, economic performances etc. our survey also consisted of two sets of candidate choice experiments(ten choices per set). In each turn the respondents had to choose between two hypothetical candidates. Each candidate was described with four randomly assigned attributes: their party affiliation, two relevant policy positions, and then either a neutral or an undemocratic position. In the latter case, a candidate was described as supporting a measure that violates one of three key tenets of a liberal democracy: electoral fairness, checks and balances and civil liberties. The respond ents were then asked two questions about the candidates: which candidate they would prefer and how likely they would vote for the chosen candidate in the next election. In total, each respondent chose from a total of 20 different candidate choice scenarios, which, in total, resulted in 200,020 experiment combinations. These experiments allowed us to isolate the effect of a candidate’s attempt to subvert democracy on their electoral prospects. Experiments The candidate choice experiments allowed us to see how a candidate’s attempt to subvert democracy affects their electoral prospects. Example of the candidate profile from the Germany survey 13

How democratically elected politicians get away with autocratizing their country POLICY POSITIONS When designing political candidates’ policy positions, we followed several criteria: Positions should be unambiguously classifiable along a single dimension(e.g., economic left-right and pro-/anti-immigration); the anticipated distribution of policies most preferred by respondents should span all policy positions and reflect the underlying distribution of preferences in the electorate; and the selection and content of the issue areas should reflect the most prominent contentious public policy issues. 2 Taxes and education Raise personal income taxes and use the revenue to increase education budget Keep personal income taxes and education budget as they are Cut personal income taxes and reduce education budget accordingly Taxes and the environment Raise taxes on coal and gas to lower the price of solar and wind energy Lower taxes on coal and gas even if it results in lesser use of solar and wind energy Immigration Allow immigration regardless of the country of origin Allow immigration from the EU, but only family-based immigration from outside of the EU Ban immigration from outside of the EU Ban all immigration regardless of the country of origin The use of minority languages 3 Only[MAJORITY_LANGUAGE] should be used when communicating with state offices [MINORITY_LANGUAGE] speakers should have the right to use their language when communicating with state offices EU Same-sex couples’ rights Same-sex couples should have the right to marry and adopt children Same-sex couples should have the right to marry, but not adopt children Same-sex couples should not have the right to marry or adopt children National versus EU-level policy on defense 4 Defense policy for [COUNTRY NAME] should be decided at the EU level Defense policy for [COUNTRY NAME] should be decided jointly by[COUNTRY ADJECTIVE] and EU institutions Defense policy for [COUNTRY NAME] should be decided by[COUNTRY ADJECTIVE] institutions alone RUS Relations with the EU and Russia 5 We should have closer relations with the EU and distance ourselves from Russia We should have close relations with both the EU and Russia 2 Accordingly, some issue areas such as immigration, EU-Russia, language, and defense policy do not apply to all seven countries under investigation. 3 Only used in Estonia, Serbia, Spain and Ukraine. 4 Only used in EU member countries. 5 Only used in non-EU member countries: Serbia and Ukraine. 14

IDENTITY, PARTISANSHIP, POLARIZATION UNDEMOCRATIC POSITIONS When designing the candidates’ undemocratic positions, we applied the following criteria: They must capture violations of key democratic principles; they must approximate practices that have been used or attempted by governments and politicians to subvert the democratic process throughout Europe; they are adopted with the intent of staying in or acquiring power(and not for their own sake); and they range from positions that are legal but undemocratic to those that are both illegal and undemocratic. This mirrors the incremental nature of democratic backsliding. Moreover, they must be presented in a manner that avoids conspicuousness or normatively leading language(“undemocratic”,“unconstitutional”,“violation”, etc.). We wanted our respondents to decide for themselves whether or not a position violates a democratic principle and thus avoid the objection that respondents are punishing undemocratic positions due to the manner in which they are presented instead of their content. In our selection of undemocratic positions, we followed classic democratization concepts by Dahl(1971, 1989) and others(Linz and Stepan 1996), and thus selected undemocratic positions in relation to three key tenets of a liberal democracy: electoral fairness, checks and balances and civil liberties. Electoral fairness Checks and balances “We should direct infrastructure spending to districts that voted for our party.” Opposition version:“If we win, we should direct infrastructure spending to districts that voted for our party.” Key concept: voter autonomy, the government should not abuse public resources for electoral purposes “Laws should be passed without parliamentary debate if criticized by the opposition.” Only plausible for candidates affiliated with the government Key concept: legislative checks on the executive “Encouraged their supporters to violently disrupt campaign rallies of their political opponents.” Key concept: freedom of assembly “Said the government should discipline judges who publicly criticize it.” Key concept: judicial independence Civil liberties “Said the government should prosecute journalists who accuse the prime minister/president of corruption” Only plausible for candidates affiliated with the government Key concept: freedom of speech “Supported a ban on foreign funding for domestic non-governmental organizations critical of the government” “Supported a ban on foreign funding for labour unions critical of the government” Key concept: freedom of association, speech “Said Muslims should not be allowed to pray during their breaks at work.” Key concept: freedom of religion “Supported a proposal for the government to monitor politically critical posts on social media.” Key concept: freedom of speech, right to privacy By randomly assigning undemocratic positions, we were able to compare the share of votes received by undemocratic candidates to that of democratic but otherwise identical candidates. A decline in an undemocratic candidate’s vote share is in effect a metric for the punishment that voters are willing to dispense in defense of democracy. Hence, this research design helped us to answer three key questions about the role of ordinary citizens in democratic backsliding: 1. Do they value and cherish democracy enough to punish politicians who violate democratic rules? 2. Are they willing to punish politicians who violate democratic rules even when this requires voting against their own political interests or party loyalties? 3. Do citizens tolerate undemocratic behavior in polarized societies more or is this behavior not connected to the degree of polarization? 15

How democratically elected politicians get away with autocratizing their country 1- CANDIDATES 1 Respondents were presented with a scenario featuring two candidates in a table format, each described by a different party label and several policy positions. All party labels and policy positions were assigned at random. 2- PLATFORMS 2 The candidates were characterized by several relevant economic and social policy positions and, crucially, by a position that either complied with or violated democratic principles. Candidate A Candidate B 3- CHOICES Respondents were then asked to choose between the two candidates and how likely they are to turn out to vote for their preferred candidate. Candidate A Candidate B ? 3 200,020 generated experiments 5 5- RESULTS A comparison of the vote shares received by candidates who comply with democratic principles and those who violate them in effect reveals the electoral punishment that voters are willing to dispense in defense of democracy. More detailed comparisons reveal the trade-offs that voters are willing to make between democratic principles and other political considerations, like partisan loyalty and preferences for specific policies. 6 4- SCENARIOS Each respondent chose from a total of 20 different scenarios. 4 CANDIDATE CHOICE EXPERIMENTS VERSUS ATTITUDINAL QUESTIONS ABOUT SUPPORT FOR DEMOCRACY When political scientists measure the public’s commitment to democracy around the world, their measurement is typically based on questions that ask directly about attitudes toward democracy as a governing system. However, recent developments in many countries across the world raise questions about the relevance of such conventional measures of support for democracy. Many countries experienced democratic backsliding while showing solid commitment to democracy when measured by conventional, attitudinal questions about a public’s support for democracy. This might be because, by asking about support for democracy directly, conventional measures invite socially desirable, politically correct responses. By contrast, in the candidate choice experiments, a candidate’s undemocratic position appears as just one among several characteristics – thus mirroring the nature of choices and trade-offs that citizens face when participating in real-world elections. According to this methodology, respondents’ support for democracy is measured not by professing support for abstract democratic ideals but by their readiness to put democratic principles above partisan interests in concrete, real-world settings. The candidate choice experiments are therefore a more realistic assessment of a public’s support for democracy: they account for the fact that in real-world elections citizens often face a trade-off between democratic principles and other valid political interests. 6 A comparison of the vote shares received by candidates who comply with democratic principles and those who violate them allows us to infer the strength of citizens’ commitment to democratic principles. Specifically, contests between two candidates that comply with democratic principles result, by design, with each candidate receiving about 50% of the vote. By contrast, a contest between a candidate who violates democratic principles and one who complies with them will typically result in a smaller vote share for the former. The magnitude of the average decline of that candidate’s vote share from the 50% benchmark in effect reveals the electoral punishment that voters are willing to dispense in defense of democracy. 16

COMPARATIVE PART The following chapter consists of results to questions that were designed to assess respondents’ understanding of democracy. Its purpose was to measure the respondents’ knowledge and support for democracy prior to the candidate choice experiments. It further analyses tendencies among the electorate across all countries to punish or reward politicians for undemocratic positions.

How democratically elected politicians get away with autocratizing their country KNOWLEDGE ABOUT AND SUPPORT OF DEMOCRACY DID THE RESPONDENTS REALLY KNOW IF A CANDIDATE’S POSITION WAS UNDEMOCRATIC? At the beginning of the survey, respondents were faced with a number of statements, with some statements consistent and others violating key democratic principles. Respondents were asked to rank these statements on a scale from 1(not at all democratic) to 10(completely democratic). These statements were created in such a way that they intentionally mirrored the undemocratic positions held by candidates in our candidate choice experiments; they thus allow us to examine our respondents’ understanding of democracy. Our findings reveal that citizens in the seven countries under investigation have a solid understanding of what democracy is and what it is not. Indeed, on average, the citizens of all 7 countries rank items which experts would classify as consistent with democratic principles as more democratic than items that experts would classify as not consistent with democratic principles. If we define a“democratically competent” respondent as one who, on average, rates the four democratic items as more democratic than the ten undemocratic items present in the sur vey, then democratic competence varies from 77% in Ukraine to 85% in Sweden. The most democratically rated item inconsistent with democratic principles is“the military removed a corrupt president.” This holds for all countries in the sample, with, on an already mentioned 10-point scale, the mean score ranging from 4.3 in Sweden to 5.2 in Poland. Interestingly, the percentage of those who find this statement very democratic ranges from 13% in Sweden to 22% in Poland. This finding also shows how much voters dislike corruption and suggests that many citizens are willing to tolerate an extra-constitutional interference by the military if it reduces corruption. Democratic competence How high is the share of people who are democratically competent? Sweden Germany Estonia Serbia Spain Poland Ukraine 20% 40% 60% 85% 82% 81% 80% 79% 78% 77% 80% “How democratic do you find the practice of the military removing a corrupt president?” Share of respondents rating this statement with a 9 or 10 on the scale from 1 to 10. Poland Serbia Spain Germany Estonia Ukraine Sweden 22% 21% 17% 16% 15% 15% 13% 5% 10% 15% 20% 18

IDENTITY, PARTISANSHIP, POLARIZATION DO OUR RESPONDENTS VALUE DEMOCRACY? One of the core goals of this study is to investigate for which economic, social, or identity-based interests citizens are willing to trade off democratic principles. Hence, we first had to establish how important democracy is to the people. To that end, the beginning of our study aims at assessing public support for democratic principles in the countries under investigation. Democracy, as a form of government, is still favored by the majority of their citizens. For the respondents in all seven countries under investigation, it is very important to live in a country that is governed democratically. Even though there are variations in this respect, they are not significant. Estonia records the lowest mean score with 8.1, while Sweden the highest with 8.8. 7 The support for democracy is also recognizable in the responses to other democracy-related questions. In all seven countries, a great majority finds having a democratic political system either“very good” or“fairly good.” Yet, we could observe a West/East divide. Most respondents in Sweden, Germany and Spain find having a democratic political system “very good”, while in Serbia, Ukraine, Estonia and Poland “fairly good” had a higher response rate. “How important is it for you to live in a country that is governed democratically?” Not at all important Mean values Absolutely important Sweden 8.8 Germany 8.5 Ukraine 8.5 Spain 8.4 Poland 8.3 Serbia 8.2 Estonia 8.1 1 2 3 4 5 6 7 8 9 10 With respect to other questions, one could observe more variation. Apart from Ukraine, in all of the countries under observation a great majority either chose“strongly agree” or “agree” when asked if democracy is better than any other form of government. With the exceptions of Ukraine and Estonia, in all countries a great majority thinks that having a strong leader who“does not have to bother with parliament or elections” is either“very bad” or“fairly bad”(most respondents see it as “very bad” or“fairly bad” in Sweden and Germany). 8 Against this background it is also noteworthy that citizens from Estonia – an EU member state – appreciate democracy comparatively little. There is strong opposition to army rule across the board. In all countries, the majority thinks that having the army rule is “fairly bad” or“very bad”(with Estonia and Germany being the most opposed, whereas Ukraine is the most accepting). 9 The political developments in Ukraine during the last decade, which led to growing insecurities, may have resulted in the acceptance of a higher level of importance for the army and the existence of strong leaders. “Having a democratic political system is either‘very good’ or‘fairly good’” “Democracy is better than any other form of government” Sweden 65% 29% 93% Sweden 50% 43% 92% Spain 54% 36% 90% Germany 47% 45% 92% Germany 47% 39% 87% Spain 44% 45% 89% Estonia 33% 49% 82% Poland 33% 56% 89% Ukraine 34% 46% 80% Serbia 24% 61% 85% Serbia 29% 49% 78% Estonia 23% 61% 84% Poland 36% 40% 77% 20% 40% very good 60% 80% fairly good Due to rounding, totals may not correspond with the sum of the separate figures. Ukraine 15% 38% 20% 40% strongly agree 53% 60% agree 80% 7 On a scale where 1 means it is“not at all important” and 10 means“absolutely important.” 8 61% in Estonia, 82% in Germany, 75% in Poland, 72% in Serbia, 77% in Spain, 85% in Sweden and 50% in Ukraine. 9 90% in Estonia, 87% in Germany, 81% in Poland, 80% in Serbia, 84% in Spain, 88% in Sweden and 76% in Ukraine. 19

How democratically elected politicians get away with autocratizing their country DEMOCRATIC VIOLATIONS, ELECTORAL PUNISHMENT AND VOTE LOSS COMPENSATIONS ARE VOTERS WILLING TO PUNISH UNDEMOCRATIC BEHAVIOR? The previous sections showed that voters in all seven countries under investigation understand what democracy is and gen erally cherish it – even if only as an abstract principle. Our experiments further examined whether citizens value democratic principles enough to punish political candidates who violate them by deciding to vote for another candidate. In the following, we define“punishment” as a loss of vote share that politicians might experience after adopting undemocratic rhetoric or behavior. 10 The answer is yes, but not very much. Indeed, respondents in the seven countries showed a willingness to stop voting for candidates who undermine democracy. However, the average punishment across the entire sample is a loss of only -7.8% of the overall vote share for the candidate who adopts an undemocratic position(when they compete against a candidate who remains democratic). 11 There are, however, significant differences in citizens’ willingness to punish candidates who undermine democracy across the seven countries. Sweden and Poland stand out as punishing the most (-10.0% and-9.2% of overall vote share loss, respectively), which is at least somewhat unexpected bearing in mind the successful attacks on the Polish democracy by the ruling party PiS in the last few years. Voters in Spain and Ukraine punish the least(-5.9% and-7.1% of the overall vote share loss, respectively); while this was to be expected in the case of a flawed democracy such as Ukraine, the result is rather surprising in the case of a consolidated democracy such as Spain. The remaining countries rank somewhere between these levels of punishment. Average punishment for undermining democracy Loss of vote share Sweden Poland Serbia Estonia Germany Ukraine Spain -7.8% This is the average punishment corresponding to the share of voters willing to defect from an otherwise favored candidate once they adopt an undemocratic position. In the following sections, we make use of regression analyses to calculate the level of punishments while taking into consideration a range of other aspects(e.g., whether the candidate was the respondent’s co-partisan or whether the candidate proposed the respondent’s favorite economic policy). Punishment punishment ‘Punishment’ is defined as a loss of -2% vote share that a politician experiences for violating -4% democratic principles. -6% -5.9% -7.9% -7.8% -7.5% -7.1% -8% -10.0% -9.2% 10 Note that the design of our experiment implies that when two candidates who adhere to democratic principles compete against each other, this results in a tie – each candidate receives 50% of the total vote. In turn, we will say that a candidate was punished for adopting an undemocratic position whenever his vote share drops below 50%. 11 According to a similar survey conducted in the US, the average punishment there is a loss of-11.7% of the overall vote share(Graham and Svolik 2020). 20

IDENTITY, PARTISANSHIP, POLARIZATION Electoral fairness Voters value the electoral component of a liberal democracy more than the civil liberties and checks and balances. WHICH UNDEMOCRATIC POSITIONS ARE VOTERS MOST WILLING AND LEAST WILLING TO PUNISH? On average, all undemocratic positions are punished, ranging from a-5.0% vote share loss for“banning foreign union funding” to a-12.2% vote share loss for “violently disrupting rallies”; in contrast, there does not appear to be any reward in general for undemocratic positions. Hence, regarding the key elements of a liberal democracy, electoral fairness (-9.8% vote share loss) appears to be valued more highly by the voters in the seven countries than civil liberties(-7.1% vote share loss) and checks and balances (-6.8% vote share loss). On the one hand, on average,“violently disrupting opponents’ rallies” and“prosecuting journalists” are punished the most, followed by“banning prayers for Muslims”, and“the government disciplining judges.” Among other things, these findings show that despite the recent rise of support for parties deploying anti-Muslim stances, in general, the populations in the countries under investigation cherish religious tolerance. On the other hand,“banning foreign funding for labor unions” and“passing laws without parliamentary debate” are punished the least, followed by“banning foreign NGO funding” and“monitoring social media.” Hence, the liberal component appears to be less cherished than the electoral one, which might explain the successful proliferation of defective/ flawed democracies. In addition to these more general observations, we were also able to identify country-specific particularities. For example, in Spain,“prosecuting journalists” is one of the two least punishable undemocratic positions, and this is also the case in Ukraine with“government disciplining judges.” As already noted, there is a particularly low acknowledgment for the importance of the oversight role of the parliament in the form of parliamentary debates. In fact, this is the least punishable undemocratic position in five countries, with Poland and Serbia being an exception. One explanation could be that the adoption of laws with a fasttrack procedure has been a pressing issue in both Poland and Serbia, which is why the electorate appears to be more aware of its undemocratic nature. Interestingly, while“partisan infrastructure spending” is one of the two least punishable undemocratic positions in Serbia, in Spain and Ukraine, the opposite is the case as this undemocratic position is among the two most punishable. Average punishment by key elements of liberal democracy Electoral fairness Civil liberties Checks and balances punishment -9.8% -7.1% -6.8% 21

How democratically elected politicians get away with autocratizing their country Average punishment for candidates by undemocratic positions All figures in% All countries average Estonia Germany Poland Violently disrupt rallies -12.2-11.5-12.4-11.2 Prosecute journalists -10.0-13.0-8.4-18.2 Ban prayers for Muslims -8.6-7.4-11.1-7.4 Discipline judges -8.2-8.4-7.1-14.1 Partisan infrastructure spending -7.3-6.8-6.3-7.6 Monitor social media -6.0-5.5-6.5-7.6 Ban foreign NGO funding Pass laws without parliamentary debate Ban foreign union funding -5.6-8.6-6.1-4.1 -5.4-5.4-3.3-14.5 -5.0-6.4-4.4-4.6 Serbia -11.8 -12.6 -6.3 -8.1 -5.1 -5.1 -6.3 -9.7 -4.8 Spain -11.9 -3.0 -5.7 -5.7 -7.5 -2.9 -5.5 -2.5 -6.1 Sweden -17.3 -14.5 -10.0 -13.7 -10.4 -8.1 -5.1 -4.9 -6.3 Ukraine -10.6 -5.8 -4.2 -2.8 -8.3 -6.6 -4.2 -4.0 -3.8 WHAT INFLUENCE DOES PARTISANSHIP PLAY? When we analyze the voting behavior for those respondents who favor the party of the candidate who advocates an undemocratic position and those who do not we found that voters are“partisans first” and“democrats second.” Most voters forgive politicians who violate democratic norms if they represent the voters’ preferred party. In other words, whenever a candidate with an undemocratic position comes from the voter’s favorite party, they will usually be compensated so that in the end they will not be punished, but rather rewarded. While the lowest compensation rate was found for Sweden, where a candidate who belongs to the respondent’s favorite party is rewarded with+21.6% in vote shares, the highest was found for Poland, where the reward is a plus of+29.0%. In the context of our experiment, this means that an undemocratic candidate of a respondent’s favorite party will typically defeat a democratic candidate from a different party. Party loyalty 21.7% Final reward 19.9% 18.0% 15.9% 15.6% 15.4% 11.5% +20% +15% +10% compensation 27.6% 22.9% 23.5% 23.7% 25.1% 29.1% +5% 21.5% 0% punishment -5.9% Spain -9.2% -7.1% Poland Ukraine Average punishment -7.8% -7.9% -7.5% Estonia Serbia Germany Compensation by voters’ party loyalty -10.0% Sweden -5% -10% 22

IDENTITY, PARTISANSHIP, POLARIZATION Party loyalty The two most loyal party supporters(by party) per country Estonia EKRE ID Isamaa EPP 27.8% 23.3% Germany AfD ID CDU/CSU EPP 25.9% 25.9% WHICH VOTERS ARE LEAST LIKELY TO PUNISH THEIR FAVORITE PARTY’S CANDIDATE? Voters who are the most loyal to their supported party’s candidate – even if those show undemocratic behavior – are those who favor right-wing populist parties: EKRE(Estonia), AfD(Germany), PiS(Poland), SNS and Dosta je Bilo(Serbia), Vox(Spain) and the SD(Sweden). We also found comparatively many voters of center-right parties to be“partisans first”: Isamaa (Estonia), CDU/CSU(Germany), PO(Poland), PP(Spain) and Servant of the People(Ukraine). 12 Notably, we also found that supporters of the Swedish Social democrats are the most partisan party supporters in that country. Poland Serbia Spain Sweden Ukraine PO 31.1% EPP IS THERE A DIFFERENCE IN PUNISHMENT PiS 29.3% RATES BETWEEN THE SUPPORTERS OF ECR DIFFERENT PARTY FAMILIES? SNS EPP Dosta je bilo ECR 26.0% 23.4% Vox ECR ERC Greens/EFA 32.5% 32.0% Socialdemokraterna S&D Sverigedemokraterna ECR 24.1% 22.9% Servant of the People ALDE 26.0% European Solidarity EPP(observer) 25.7% 0% 10% 20% 30% In general, voters of key parties in all investigated countries were inclined to punish undemocratic politicians. Among them, the voters of the right-wing Polish PiS demonstrated the least willingness to punish undemocratic behavior(-3.4% vote share loss), while the voters of Swedish Greens the most willingness (-20.9% vote share loss). We furthermore identified that across the entire sample, in general, voters of the far-right or rightwing parties are the least willing to punish undemocratic behavior. Among the former, the highest punishment rate is -7.3% in the case of Serbian Dveri voters. Our experiment also shows that the voters of the far-left or left-wing parties do punish undemocratic behavior, on average even more than the center-right or center-left parties. Their average punishment rate ranges from-7.2% in the case of Spanish Unidas Podemos voters to a-13.4% punishment rate in the case of Swedish Vänsterpartiet voters. This suggests that the far-left parties are far more willing to punish undemocratic behavior of a candidate and are hence not contributing to a backsliding of democracy as is the case for right-wing or far-right parties. There is yet another group that is consequently tolerating of undemocratic behavior across all Western European countries: non-voters hardly punish undemocratic candidates, ranging from a mere-2.6% punishing rate in Sweden, over-4.4 in Germany and-0.5% in Spain. 12 SNS of Serbia is actually affiliated with the center-right European People’s Party, but can rather be considered a right-wing populist party. 23

How democratically elected politicians get away with autocratizing their country WHAT KIND OF UNDEMOCRATIC BEHAVIOR DO THE PARTY SUPPORTERS PUNISH DIFFERENTLY? It should also be noted at this point that a willingness to punish does not apply to each violation of democratic principles to the same degree. Indeed, the voters of some parties actually reward the candidate who advocates particular undemocratic positions. Besides the already mentioned voters of the PiS, the voters of other parties – the far-right SNS, the Spanish centerleft PSOE and the Ukrainian center-right Servant of the People – also reward disciplining of judges who criticize the government. The voters of these parties(bar the SNS) reward the prosecution of journalists as well. What is more, passing laws without parliamentary debate seems to be an undemocratic position that is the most acceptable across our sample, as the voters of five parties – the CDU/CSU, the PiS, Podemos, Servant of the People and the center-left Socialdemokraterna – reward it. The voters of only one party – the Swedish far-right SD – reward a ban on prayers for Muslims during their work break with their voting behavior. The same holds for PiS voters with respect to the ban on foreign NGO funding. When we combine our findings on party loyalty and on the punishing behavior of the different party supporters, we can see that the parties associated with politicians who have lately been undermining democracy – Jarosław Kaczyński in Poland and Aleksandar Vučić in Serbia – are not to suffer in the elections amid such a behavior of their leadership. Besides being the most loyal voters in Serbia, the voters of Vučić’s SNS do not show a strong willingness to punish violations of democratic principles either, as the punishing rate is only-4.1% vote share loss. Generally, we found that right-wing party supporters are the most likely to let an autocratization of their country happen. But we can also see an alarming degree of tolerance of undemocratic behavior among voters of parties that have traditionally been considered the backbone of democracy in Europe and that traditionally aim to fulfill a gatekeeping-function (Levitsky and Ziblatt 2018) to prevent an autocratization of their respective country. Noteworthy examples of this tendency in consolidated Western European democracies is the behavior of voters who prefer the center-right CDU/CSU(Germany) and Partido Popular(Spain) as well as the center-left PSOE(Spain). Willingness to reward undemocratic behaviour by parties Selected examples from the political spectrum throughout Europe Podemos (ESP) Socialdemokraterna (SWE) PSOE (ESP) CDU/CSU (GER) Serv. People (UKR) SD (SWE) SNS (SRB) PiS (POL) Discipline judges Prosecute journalists Pass laws without parliamentary debate Ban prayers for Muslims Ban foreign NGO funding 11.2% 16.4% 10.7% 19.8% 9.3% 7.5% 12.1% 18.6% Far-Left Center-Left Center-Right 14.8% 29.0% 20.8% 26.2% 7.1% 4.3% Far-Right 24

IDENTITY, PARTISANSHIP, POLARIZATION POLICY PREFERENCES Voters in all of the countries under investigation are willing to punish candidates who adopt undemocratic positions with a loss of vote share, albeit to varying degrees, as we discussed in the preceding section. Nevertheless, politicians who violate democratic principles continue to enjoy electoral support. The experimental part of our survey explored this seemingly contradictory relationship between voters and undemocratic political candidates in greater depth. In this section, we examine the possibility of undemocratic political candidates compensating for their lost votes by offering voters policies aligned with their preferences(e.g., socio-economic or identity-related policies). We also examine whether these trends in vote compensation are further reinforced when the political candidate represents the voter’s preferred political party. Identity-related interests Our results indicate that voters are most forgiving of undemocratic politicians when they appear in combination with identity-based policy preferences. In our experiment we consider the following topics to be primarily identity-related policy 13 preferences: rights for same-sex couples, immigration, language and the EU vs. Russia foreign policy-orientation. The issue of environmental policies(the taxing of renewables and fuels) could also be seen as a social identity-related issue rather than a question that concerns socio-economic interests only. We found that voters would disregard democratic violations if their preferred identity-based policies were legislated. We refer to this phenomenon as“the identity-democracy trade-off.” As a result, we can conclude that identity policies protect politicians from losing votes, even when they violate democratic principles. Indeed, voters in all seven countries tolerate undemocratic behavior when it occurs in relation to identity-based interests(apart from the immigration issue in Poland). The issue that is on average the most salient for voters concerns“rights of same-sex couples.” The latter appears even more salient than immigration. As outlined in more detail below, in exSame-sex couples’ rights: shift in punishment when presented with favored policy Voters disregard democratic violations in exchange for their preferred policy regarding the rights of same-sex couples How to read the graph 1 -13.3% 2 +21.5% 8.2% 3 1 The regular punishment of undemocratic behaviour among left-leaning voters equals -13.3%. 2 In this example, +21.5% is the absolut compensation rate that a candidate gets when they offer voters their favorite policy regarding same-sex couples‘ rights. 3 The relative compensation(or in some cases, punishment) is the final vote share increase or electoral reward the candidate receives when both the initial punishment and the overall compensation are taken into account. punishment Germany reward -9.6% +16.7% 7.1% +8.6% -3.6% 5.0% -12%-8%-4% 0%+4%+8%+12% Left Right punishment Sweden reward -12.1% +19.0% 6.9% punishment Spain reward -7.0% Left +17.8% 10.8% Left +5.6% 4.5% -1.1% -12%-8%-4% 0%+4%+8%+12% Right +0.3% -2.2% -2.5% -12%-8%-4% 0%+4%+8%+12% Right 13 In our study, we examine immigration as a social identity issue based on the way politicians, particularly right-wing populists, have framed immigration policies during their campaigns in recent years. Thus, based on our results from the candidate-choice experiments, we aimed to infer how such political framing on immigration policies has influenced electoral behaviour. However, we also acknowledge the possibility that voters’ preferences for certain immigration policies are driven by economic grievances(e.g. Halikiopoulou& Vlandas 2020, Stockemer, Halikiopoulou& Vlandas 2021, Halikiopoulou & Vlandas 2022, Rydgren, 2008, Mayda 2006). Similarly, we postulate that motivations related to social identity and group belonging are of significant 25 importance for explaining voters’ preferences on environmental policies(e.g. Brieger 2018).

How democratically elected politicians get away with autocratizing their country change for a preferred policy regarding rights of same-sex couples, voters are often very forgiving of undemocratic politicians and even reward them with extra share votes. For instance, in Estonia, a politician who violates democratic norms would generally lose-7.8% of their vote share. They can however compensate for this initial loss when they proposed the voters’ preferred policy on the rights of same-sex couples: this scenario results in a+17.8% vote share increase, which means the candidate ends up with a total of a+10.0% reward in vote shares. This is the highest increase in vote share observed in relation to policy preferences on the rights of same-sex couple. On the other end of the spectrum, in Poland, a politician who violates democratic principles would initially lose -9.3% of their vote share, which they could compensate by adopting voters’ preferred policies on the rights of same-sex couples by+11.2%, resulting in a final reward of+2.0% in total vote share increase. This is the lowest compensation rate observed across the sample with regards to this policy. The same applies when looking at voter’s preferred immigration policy, albeit with a lower vote share increase, with the lowest observed in Germany and the highest in Sweden. In Germany, an undemocratic politician would lose-7.5% of their vote share for violating democratic norms. However, once the candidate proposed the voters’ preferred policy on immigration, they could compensate with a+10.2% vote share increase, which would leave them with a total of a+2.7% reward in vote shares. We observe the highest reward in Sweden:+4.2%(after a loss of vote share of-10.0% for democratic violations and a compensation of+14.2%). With the exception of Serbia, the language policies are generally almost as important as the rights for same-sex couples. In Estonia, Spain, Serbia and Ukraine voters tolerate a candidate’s undemocratic behavior if they also propose their preferred language policies. We observe the lowest reward in vote share in Serbia, a plus of+1.3%, after a punishment rate for undemocratic behavior of-7.9% and a compensation rate of+9.2%. punishment Estonia reward punishment Poland reward -13.3% -13.0% +21.5% 8.2% Left +11.0% -2.0% Left -5.6% +18.9% 13.3% Right -12%-8%-4% 0%+4%+8%+12% punishment Ukraine reward -8.2% +12.3% 4.1% Left +12.8% -4.3% 8.5% -12%-8%-4% 0%+4%+8%+12% punishment Serbia reward -8.9% +5.7% -3.2% Right Left +14.9% -6.1% 8.8% -12%-8%-4% 0%+4%+8%+12% Right +13.7% -8.1% 5.6% -12%-8%-4% 0%+4%+8%+12% Right 26

IDENTITY, PARTISANSHIP, POLARIZATION Party loyalty Most voters forgive politicians who violate democratic norms if they represent the voters’ preferred party. The highest reward in vote share pertaining to language could be observed in the case of Ukraine:+6.8%. In Serbia and Ukraine, we were able to detect similar patterns with regards to the voters’ preferred policies on foreign policy. The country where environmental policy preferences influence voters’ readiness to overlook an undemocratic candidate’s undemocratic behavior most is Germany. While voters would initially punish candidates who violate democratic norms with a loss of-7.5% of the vote share, they would also forgive such candidates in exchange for their most preferred environmental policy. This results in an increase of+11.1%, which would ultimately amount to+3.6% vote share reward. To a lesser extent, similar compensation rates can be observed in Sweden as well. Socio-economic preferences Regarding the socio-economic interests, the results are not as clear as regarding identity-related issues, especially since there are noteworthy differences between the countries. For instance, only in Germany and Estonia politicians were able to compensate for democratic violations once they offered their constituencies their favorite policies on tax and education. In Germany, such candidates would compensate with an in crease in vote share of+8.1%, which in combination with the initial loss of vote share for being undemocratic, would imply a+0.6% final vote share increase. Similarly, in Estonia, the reward in vote shares stands at+0.8%. When it comes to environmental policies, we can observe that in the Eastern European countries, undemocratic candidates were either not able to get any vote compensations(even when they offer voters their favorite policies on the environment) or they get rewarded with a minimal increase in vote share(below+1%). In contrast, in Germany, once undemocratic candidates offer their voters their favorite environmental policy, they can compensate for the initial vote share loss of-7.5% with an increase of+11.1%, which would ultimately amount to+3.6% in vote shares. Lastly, the least significant policy area when it comes to forgiving undemocratic politicians is defense. Only in Estonia, the behavior would be ultimately rewarded with a plus of +0.1% in vote share. ARE SUPPORTERS OF A CERTAIN POLICY INTEREST MORE FORGIVING OF UNDEMOCRATIC BEHAVIOR THAN OTHERS? In the previous chapter we depicted the average willingness of the electorate of the seven countries under investigation to forgive undemocratic behavior if the candidate advocates certain favorite policies. However, the question arises whether this tendency can be observed across the entire political spectrum, i.e., whether one can observe differences between left-leaning and right-leaning voters. In some cases, a policy issue may become so politicized that policy supporters feel very strongly about it. It is therefore useful to look at how the supporters of a certain policy issue would forgive or not forgive the undemocratic behavior of a candidate. Taxes and education Across all countries, the rate for the loss of votes of a candidate induced by economically left-standing voters is higher than the average loss-of-vote-share-rate in the country. Citizens who want to see their income tax decrease together with the education budget are significantly more forgiving of a candidate’s undemocratic behavior than their country’s average. An exception is Ukraine, where the tendency is the same, but the difference is not as significant as in the other countries. Moreover, in all countries we found that generally people who would want the education budget raised at the cost of higher income taxes tolerate undemocratic behavior less than people who are economically positioned on the right or libertarian end of the political spectrum. Interestingly, when presented with a candidate who at the same time also shared their respective favorite policy position on taxes and education, left-lean27

How democratically elected politicians get away with autocratizing their country Interests and punishment/rewards 14 All figures in% Average punishment Estonia -7.8 Partisan +15.8 Germany -7.5 +15.4 Poland -9.2 +20.0 Serbia -7.9 +15.6 Spain -5.9 +21.0 Sweden -10.0 +11.5 Ukraine -7.1 +18.0 Example In Estonia, a political candidate who violates democratic norms will on average be punished with -7.8%* of the vote share. Average punishment -7.8 Partisan com re p w en a s r a d tion +15.8 However, if the candidate comes from the voter's favorite party, the candidate is able to compensate for their initial loss of vote share by a total of+15.8% vote share i r n e c w re a a rd se . . T T h h is is p p re re ve ve n n ts ts th th e e m m fr f o ro m m losing the election in spite of undemocratic positioning. * this is the case when the candidate is not from the voter's favorite party and they have not proposed any of the voter's favorite policies Education +0.8 Environment -0.4 Same-sex couples‘ rights +10.0 Immigration +3.7 EU-Russia n/a Language Defense policy +5.5 +0.1 +0.6 +3.6 +5.8 +2.7 n/a+0.0 -0.5 -3.4 -0.9 +2.0 -1.7 n/a n/a -2.4 -1.5 +0.1 +5.3 n/a +4.8 +1.3 n/a -0.2-0.5 +8.4 +3.1 n/a +5.4 n/a -1.8 +1.8 +5.8 +4.2 n/a n/a -1.8 -2.0 +0.2 +5.1 n/a +9.3+6.8 n/a punishment -+ c re o w m a p rd ensation ing voters in Estonia, Serbia, Spain and Ukraine deviated more from their original punishing rate towards forgiving undemocratic behavior. These trends might show the importance of the policy issue for this voter group in these four countries. However, across all countries, when presented with their favorite economic policy, voters of the right punished undemocratic candidates less than left-leaning voters. Most noticeably, in Sweden and Germany, economically right-wing voters even rewarded undemocratic candidates as a trade-off for their preferred policy. In Sweden, right-wing voters compensated candidates with a plus of+12.0%, which decreased to a final reward rate of+6.4% in relation to the punishment rate of this cohort(-5.6%). Similarly, in Germany, right-wing supporters would ultimately reward undemocratic politi cians in exchange for their favorite policy on education and taxes with an increase of+6.7% in vote shares. Taxes on coal and gas In all countries under investigation, in general, green-leaning voters punish undemocratic behavior more than those who prefer the use of fossil fuels. Apart from Ukraine, Spain, and to a lesser extent Estonia, there is a significant difference between the punishment rates of the two cohorts. For instance, in Sweden, the punishment rate of green-leaning voters is-12.1%, whereas the punishment rate of those who prefer“lower taxes on coal and gas even if it results in lesser use of solar and wind energy” is-5.6%. Likewise, Poland has a punishment rate of -12.6% and-7.7%, respectively. However, voters on both sides of the spectrum have a stronger tendency to punish undemocratic politicians less in exchange for their preferred economic policy on the environment. Furthermore, green-leaning voters in Germany and Sweden are more likely to forgive undemocratic politicians when they are pre 14 There is no statistically significant difference between immigration(in Sweden) and language(in Ukraine) and other social policies. 28

IDENTITY, PARTISANSHIP, POLARIZATION Same-sex couples’ rights In all seven countries under investigation, a favorite policy on same-sex couples’ rights trumps democracy. sented with their favorite economic policy on the environment. In contrast, green-leaning voters in Poland, Serbia, Estonia and Ukraine are less tolerant of violations of democratic norms. Rights for same-sex couples In all countries under investigation, voters who support an extension of rights for same-sex couples(e.g., to marry and to adopt children, a left-leaning position) generally tolerated undem ocratic behavior less than the country’s average voter. With one exception(Serbia), we also see that people who would like to exempt same-sex couples from the right to marry and adopt children(a right-leaning position) tolerate undemocratic behavior often decidedly more than the country’s average voter. Ukrainian voters also share this tendency, but there is a significantly smaller difference between those who favor exempting same-sex couples from the right to marry and adopt children and the coun try’s average voter(-6.1% versus-7.1%). Only in Serbia we see both the left and the right tolerate undemocratic behavior less than on average, with-8.9% and-8.1% respectively. Notably, this punishing behavior of socially left-leaning voters shifts drastically when voters were presented with a candidate who(besides proposing an undemocratic policy) also proposed their favorite policy regarding rights for same-sex couples: In all Western European countries(Sweden, Germany, Spain), left-leaning voters became far more tolerant of undemocratic candidates(see the graphic for visualization of the shift). Thus, we can conclude that this issue is of great importance to a number of social-leftist voters in the West. This is all the more surprising as only a relatively small share of voters would actually be affected by this policy themselves. Two aspects could explain this voting decision: on the one hand, rewarding more rights to same-sex couples is connected to a strengthening of democracy, as a minority group is given the same rights as the majority. On the other hand, it appears that voters on the social left consider the issue to be part of their group identity(they are allies of LGBT people) thereby disregarding other undemocratic behavior. In the other countries(Estonia, Poland, Serbia, Ukraine) we find that right-wing voters become as tolerant of undemocratic behavior as the left-wing voters, or even slightly more so, when presented with a candidate who proposes their preferred policy. Combined with the generally stronger tendency to forgive undemocratic behavior, right-wing voters in Estonia and Poland end up rewarding such a candidate significantly. In Estonia, voters reward undemocratic candidates in exchange for their favorite policy with a vote share increase of+18.9%. Ultimately, in relation to the initial punishment of-5.6%, the reward rate would drop to+13.3%. Similarly, in Poland, in spite of the initial punishment rate of-4.3%, right-wing supporters would ultimately reward the undemocratic candidate with a vote share increase of+8.5% in exchange for their favorite policy. This tells us that the topic is so important and salient for right-wing voters in Eastern European countries that a candidate who proposes a voter’s preferred position can easily be forgiven undemocratic behavior. Language In Estonia, voters who support the position that also minority languages should be used in state offices are more forgiving of undemocratic behavior of a candidate than the country’s average (-4.8 vs.-7.8% punishment rates in the form of vote share loss). People who favor the position that only the majority language be used when dealing with state offices are less forgiving of undemocratic behaviors of candidates than the country’s average(-9.1 vs.-7.8%). In Serbia and Spain, we see the opposite: majority language-policy supporters punish undemocratic candidates at the ballot box just about as much as the average voters, but minority language-policy supporters are more unforgiving of undemocratic behavior. Immigration Across the entire sample the left-leaning voters on immigration issue value democracy more than the right-leaning voters as they punish more than those on the right. In other words, the more left-leaning a voter is when it comes to immigration issue the readier they are to punish undemocratic behavior. When it comes to the 29

How democratically elected politicians get away with autocratizing their country willingness to forgive undemocratic behavior for a voter’s favorite policy on immigration, we see that, bar Estonia, in the other four countries in which this issue is salient the center-left voters show the strongest attachment to democratic principles. While in Spain they still reward an undemocratic candidate for advocating their favorite immigration policy – albeit with a minimal+0.2% vote share increase –, in Germany, Poland and Sweden center-left voters punish such a candidate, with a-1.4%, -4.6% and-5.6% vote share loss, respectively. It is noteworthy that we were able to observe that right-leaning voters in Germany and Estonia are more punishing of undemocratic behavior than left-leaning voters when presented with their favorite policy. Moreover, the part of the electorate that across the entire sample demonstrated the strongest willingness to forgive undemocratic behavior for its favorite immigration policy is the center-right voters’ group in Sweden, with a significant vote share increase of+12.5%. EU-Russia foreign policy orientation The issue of EU-Russia policy preferences was applicable to Serbia and Ukraine only, and our experimental analysis shows the following results: In both countries, those who are pro-Russian show less willingness to punish undemo cratic behavior. Moreover, in Serbia, there is a willingness across the entire electorate to forgive undemocratic behavior when confronted with its favorite foreign policy position. This applies in particular to those voters who want a closer relationship with Russia(+12% vote share increase). In Ukraine, the advocates of a stronger attachment to Russia would neither reward, nor punish a candidate for it, while the rest of the electorate would forgive undemocratic conduct for one’s favorite policy in this regard. This is most evidently the case with the advocates of a close relationship with both the EU and Russia(+11.2% vote share increase). Defense policy It is remarkable that the nationalist position was preferred by so few survey respondents across all EU member countries that we could not use it for our analysis. The statement“Defense policy for[COUNTRY NAME] should be decided by[COUNTRY ADJECTIVE] institutions alone” was the preferred policy for none of the respondents in Estonia, Germany and Sweden, and for only 20 respondents in Spain and 40 respondents in Poland. With the exception of Poland, respondents with the view that defense policy should be a common task of their country and the EU generally punish undemocratic behavior more than the respondents who prefer sole EU responsibility for defense. However, when presented with their preferred policy, respondents with a cooperative view become far more forgiving of undemocratic behavior across all countries than respondents who want to see defense policy in the hands of the EU alone. All in all, we can thus deduct from our policy preference analysis that voters who have more right-wing or nationalist views on social and economic issues are more tolerant of undemocratic behavior. We also found that left-leaning voters generally punish undemocratic behavior more, but are, at the same time, ready to be substantively more tolerant of undermining democracy when their favorite policy position is presented. Even then they do still punish undemocractic behavior more than right-wing voters. Immigration Swedish center-right voters show the strongest willingness to forgive undemocratic behavior for their favorite immigration policy, with a +12.5% final vote share increase. 30

IDENTITY, PARTISANSHIP, POLARIZATION WHAT ROLE DOES POLARIZATION PLAY IN AUTOCRATIZATION EFFORTS? *Agreement with policy statements To capture the difference between consensual and polarized societies, we develop an empirical measure of polarization that is based on a distributional understanding of this concept. At one extreme, a maximally polarized society is one in which citizens’ policy preferences are distributed in a U-shaped manner, with an equal number of ideal policies concentrated at opposing ends of their distribution. At the other extreme, a maximally consensual society is one where all citizens’ preferences are identical. To measure a society’s degree of polarization, we take the variance of the empirical distribution preferences in a sample and divide it by the variance of a maximally polarized society for that sample. This ratio of variances ranges from 0 to 1, with 0 corresponding to a maximally consensual society and 1 to a maximally polarized society. more important (1 10) Is it easier for politicians in polarized societies to autocratize their countries? In the previous part of the study, we illustrated the policy issues for which voters are more likely to forgive undemocratic politicians. The next part addresses the question of whether polarization contributes to the intensification of such forgiving behavior. According to scholars who study polarization (Graham and Svolik 2020; McCoy, Rahman and Somer 2018), polarization increases voters’ willingness to tolerate undemocratic behavior for policy preferences and interests. Against this background, firstly, we examined whether the policies that were presented to the survey respondents are polarizing in the seven countries under investigation. To do so, we asked the survey respondents whether they agreed(on a scale from 0 to 10) with the presented policy statements.* This allowed us to measure policy polarization based on the preferences of voters and thus identify concrete polarizing issues in the seven countries under investigation. It also enabled us to develop a more nuanced argument. In this context, we identified several patterns. The most polarizing issue in the Eastern European countries(Estonia, Poland, Serbia and How important vs. how polarizing issues are Issues displayed according to their importance to voters and degree of polarization 7.5 Environment EU-Russia 7.0 Education 6.5 Defense policy Immigration Language 6.0 5.5 Same-sex couples' rights 5.0 0.2 0.3 0.4 0.5 0.6 0.7 more polarizing (0.0- 1.0) 31

How democratically elected politicians get away with autocratizing their country Ukraine) concerns rights for same-sex couples. The West European cases under investigation are not as homogenous. While immigration is the most polarizing issue in Sweden, language plays this role in Spain and environment in Germany. In five of them – Germany, Sweden, Poland, Estonia and Ukraine- taxes and education is the least polarizing issue. The same holds true for foreign policy orientation in Serbia and defense policy in Spain. In a further step, we examined whether voters are more willing to trade-off democracy for the policy issues that we identified as polarizing. In this context, we were able to identify different types of relationships between the willingness to forgive undemocratic behavior and polarization. Firstly, the willingness to forgive undemocratic behavior for some specific interests correlates with the existence of increased polarization along some of these issues. This correlation applies to same-sex couples’ rights in all countries under investigation bar Spain; immigration in Sweden, Estonia and Spain; language in Spain, Serbia and Ukraine; environment in Sweden, Serbia and Germany; and EU-Russia foreign policy orientation in Ukraine. Secondly, the voters showed a clear willingness to forgive undemocratic behavior for specific policy preferences. Yet according to the data from our survey, we cannot define these issues as polarizing. This applies to same-sex couples’ rights in Spain; immigration in Germany; EU-Russia foreign policy orientation in Serbia; education in Germany and Estonia; defense policy in Estonia; environment in Ukraine; and language in Estonia. However, although we cannot detect an overall increased polarization, our data shows that many voters have intense preferences 15 with respect to these interests, which could explain their willingness to forgive undemocratic behavior for them. Moreover, the majority of depicted interests are identity-related interests. Such interests tend to“involve strong emotional commitments and sharp moral judg ments”(Bonikowski 2017, 189), which furthermore could prompt voters to disregard democratic principles. This effect is further strengthened by the fact that, according to our data, all these issues are salient for voters, whereas some of the depicted issues have been politicized by political actors who tried to politically benefit by exploiting them. Thirdly, renewable energies/environment is a polarizing issue in Poland, but voters did not show a willingness to forgive undemocratic behavior for it. While, when making use of the same candidate choice experiment, for the US it was proven that polarization clearly favors democratic backsliding, our analysis shows that this correlation is not as strong for European countries. In the European countries under investigation, we found that some policy issues are indeed polarizing, but do not provoke undemocratic behavior. In parallel, we found that certain voters easily trade-off some democratic standards for policy issues that are not polarizing in a country, but that they simply find important. And we also found a few issues that were polarizing in a country and that indeed made voters let democratic standards slip. These mixed findings show that polarization can make it easier for politicians to undermine democracy, but only under special circumstances. Our results also show that there are other mechanisms, namely what we called the“identity-democracy trade-off”, that are at work when democratic standards are being abolished. And often, but not always, polarization and trade-offs relating to identity issue go together. Does polarization matter? The willingness to forgive undemocratic behavior for specific interests is not always explained by an increased polarization of these interests. 15 We define having intense preferences when by assessing respective policy positions on a scale from 0(strongly oppose) to 10(strongly support) the respondent selected 0, 1, 9 and 10. 32

IDENTITY, PARTISANSHIP, POLARIZATION Political engagement The less one is politically engaged the more one punishes undemocratic behavior – except for nonvoters, who hardly punish undemocratic candidates. 33 WHO ARE THE VOTERS WHO PUNISH UNDEMOCRATIC CANDIDATES? In the previous chapter we have looked how polarization can influence a voter’s tolerance for undemocratic behavior. Now we want to see whether certain personal characteristics or backgrounds of a voter have a general impact on their tolerance of undemocratic behavior. According to research on the determinants of support for democracy(Lipset 1959; Norris 2011; Pew Research Center 2009), certain social groups place more value on democratic principles: groups that have obtained higher education, the younger population(in post-communist societies), the ones more knowledgeable and more interested in politics, and those who are more secular and trust their peers. The literature also indicates a greater ten dency for men to vote for illiberal far-right parties, whereas those with an authoritarian personality or dire eco nomic prospects are more likely to accept politicians who undermine democracy (European Election Voter Study 2019; Przeworski et al. 2012; Adorno et al. 1950). Do these expectations match with the findings from our cross-national survey? The experimental part of the study revealed that some of the aforementioned characteristics do play an important role. The assumption that women punish undemocratic behavior more than men is, for example, confirmed for all countries. In addition to it, people who are more educated, students, people who are interested in politics(except in Ukraine), those who possess a higher level of political knowledge, those who are less religious(except in Ukraine), and who lack an authoritarian personality punish undemocratic behavior more strongly as well. Interestingly we can see that people who generally have more trust in people punish on average more only in Western European countries and Estonia. There is almost no difference in the punishment rate of untrusting and trusting people in Serbia, Poland and Ukraine. We can also see some noteworthy correlation between political engagement and the tolerance towards undemocratic behavior. When asked whether citizens saw voting as a duty, as a choice or neither of those two options, we found that on average across all countries the attitude towards voting had hardly any influence on the respondents’ readiness to punish undemocratic behavior; only in Spain respondents who see voting as a duty punish less, whereas in Sweden these respondents punish more. Furthermore, we examined if there is correlation between the punishment rates for undemocratic behavior and the engagement in political activities, such as volunteering or attending rallies. Surprisingly, our results show that voters who do not engage themselves politically punish undemocratic politicians the most, whereas those who volunteer or attend rallies punish the least. To substantiate these results, we checked whether individuals who are involved in at least one political activity(which might have a positive effect on punishment) might be significantly more likely to participate in another political activity(which might have a stronger negative effect), leading to an underestimation of the punishment of participants of a certain group. Yet, we observed no strong positive correlation in this regard. On the contrary, our results show that those respondents who post on social media, for example, are the least likely to sign a petition. When it comes to a low level of willingness to punish undemocratic behavior our survey revealed that on average across all countries respondents who are

How democratically elected politicians get away with autocratizing their country Characteristics of respondents who showed more willingness to punish undemocratic politicians More trustful (not in Poland, Serbia and Ukraine) Lack authoritarian personality More political knowledge ODEMACY CR More interested in politics (not in Ukraine) Higher educated Less religious (not in Ukraine) unemployed punish undemocratic behavior less. Moreover, apart from Sweden, the recently laid off also showed a greater willingness to disregard democratic principles. Similarly, in Serbia and Sweden, voters who take care of their families, and pensioners in Ukraine, punish undemocratic politicians less compared to other social groups. We also found that many respondent characteristics have only a very weak or even no association concerning the willingness to punish undemocratic politicians(or the findings are mixed). First of all, looking at age, the young in Estonia, Sweden and Poland punish more, while in Serbia and Ukraine they punish less. In Spain and Germany, we do not see a similar linear development with age. In Spain the youngest age group(18-29 ys.) together with the mid-aged 50-65-year-olds also punish more than other age groups. And in Germany, the youngest age group(18-29 ys.) punish the least and the 50 to 65 year-olds punish the most. The mixed findings also appear to be the case with religious affiliation and political participation as well as with socio-economic attributes such as employment sector, home ownership, economic assessment of the family situation, and economic assessment of the country. Interestingly, regarding the last two characteristics there are differences among the cases. While in Poland, Serbia and Ukraine, those with a more positive economic assessment of the country punish less, in Sweden, this relationship runs in the opposite direction. Similarly, in Poland and Serbia, those with a more positive economic assessment of their family’s economic situation punish less, whereas in Estonia and Sweden, the opposite is the case. Voters in Poland with a negative assessment of their family’s economic perspective are significantly more punishing of undemocratic behavior than voters with better economic perspectives. Besides Poland, we find this in Serbia as well, albeit to a lesser degree. In other countries the voters with a satisfactory economic perspective punish undemocratic behavior slightly more than people with a more negative or a more positive assessment of their family’s economic perspective. Youth In Serbia and Ukraine, the young punish undemocratic behavior less than the average voter. 34

IDENTITY, PARTISANSHIP, POLARIZATION CONCLUSION By analyzing what drives voters to support politicians who undermine democratic principles, this study aimed to contribute to the growing literature on democratic backsliding. Through candidate choice experiments, respondents in our survey were exposed to scenarios that closely resembled real-life voting situations. As a result, we were able to investigate under which conditions and in exchange for which policy preferences voters are inclined to overlook democratic violations. Accordingly, we can draw the following conclusions about voters’ behavior and the conditions under which politicians are able to incrementally autocratize their countries: Voters do not value all elements of liberal democracy equally. They value electoral fairness more than checks and balances and civil liberties. It appears that they do not consider the latter items as indispensable as the former. Moreover, we see that voters with a greater understanding of democracy punish undemocratic politicians more. In light of this, it is imperative that future democracy support efforts educate voters accordingly. This would strengthen democracy and make it more resilient. Identity policies facilitate democratic backsliding. One of our central findings is that identity-based policies protect politicians from losing votes, even if they violate democratic principles. Against this backdrop, politicians in some countries have elevated the salience of such topics on the political agenda, thereby exacerbating divisions between their party and the opposition and facilitating a subversion of democracy as well. If democratic standards are to be upheld, they should refrain from instrumentalizing and exploiting these issues for their particular gains. Moreover, awareness of the potential for instrumentalization and exploitation of identity-based policies should be raised. Partisanship implies responsibility. Our study shows that voters are especially inclined to forgive undemocratic behavior when an undemocratic candidate is associated with the voter's favorite party. This finding indicates that politicians and political parties have a great deal of responsibility in safe-guarding democracy. They should not appeal to the voters’ lowest instincts in order to boost their chances in the elections. 35

Polarization is not the sole menace to democracy. Polarization does play an important role in the deterioration of democracy, but it is certainly not the only driver. Identifying the issues which polarize societies significantly, and hence menace democracy, is valuable for democratic resilience. Innovative methods for citizens’ participation could be a tool to counter processes of polarization, but more focus on depolarization measures will be required. Western Europe is not immune to identity-driven democratic backsliding. Not only in Eastern Europe, but also in Western Europe there is a vast potential for political actors to subvert democracy in the name of identity-based interests. This finding opposes culturally deterministic assumptions on democratic backsliding and could contribute to enhancing and nuancing future research on this topic. Voters on the right and the left play a different role when it comes to the erosion of democracy. Forgiveness of democratic violations is particularly prevalent along partisan lines, and voters of far-right parties are least inclined to punish undemocratic behavior. As a result, these voters are most conducive to undermining democratic systems. By contrast, voters on the left are generally less tolerant of democratic breaches. Yet, they value certain policies(e.g., the rights of same-sex couples) so highly that they are willing to sacrifice other democratic principles for them. Autocratization does not necessarily require polarization. Issues that are not polarizing can also be drivers of autocratization. The implications of this finding, which speaks for a more nuanced and context-specific understanding of the effects of polarization on democratic backsliding, while keeping policy salience in mind, can be used in future research on democratic backsliding. We furthermore hope that this outcome of our study will encourage future research on the phenomenon. 36

IDENTITY, PARTISANSHIP, POLARIZATION 37

COUNTRY PROFILES

IDENTITY, PARTISANSHIP, POLARIZATION Facts& Figures Population Political system Varieties of Democracy regime type Varieties of Democracy Liberal Democracy Index ranking(2021) Which issues let politicians get away with undemocratic behavior? Issues displayed according 8 to their importance to voters and degree of polarization ! Issues for which voters forgive undemocratic behavior 7 ! Education 6 Environment ! Defense policy ! Immigration ! Language more important (1 10) 5 Same-sex ! couples’ rights 4 3 0.2 0.3 0.4 0.5 0.6 more polarizing (0.0- 1.0) 81% Democratic competence of Estonian respondents Percentage of respondents who on average rate the four democratic items as more democratic than the ten undemocratic items in our survey How much are key elements of liberal democracy valued? The higher the vote loss for an undemocratic policy proposed by a political candidate, the more a certain democratic element is valued. -6.4% -5.4% -8.6% -5.5% -6.8% -8.4% -13.0% -7.4% -11.5% Ban foreign union funding Pass laws without parliamentary debate Ban foreign NGO funding Monitor social media Partisan infrastructure spending Discipline judges Ban prayers for Muslims Prosecute journalists Violently disrupt rallies -20%-15%-10%-5% 0% Average punishment across all 7 countries Min/Max punishment across all 7 countries Punishment in Estonia 39

How democratically elected politicians get away with autocratizing their country Estonian parties presented in the survey English name Estonian name European affiliation Currently governing Estonian Reform Party Eesti Reformierakond(RE) Estonian Centre Party Eesti Keskerakond(EK) Conservative People's Party of Estonia Eesti Konservatiivne Rahvaerakond(EKRE) Social Democratic Party Sotsiaaldemokraatlik Erakond(SDE) ALDE ALDE ID S&D Pro Patria Isamaa EPP KNOWLEDGE ABOUT AND SUPPORT FOR DEMOCRACY Respondents of Estonia have a good understanding of what democracy is and what it is not and they are fairly pro-democracy oriented as well. Estonians are comparatively satisfied with how democracy works in their country(ranking third just behind Sweden and Germany, albeit with some distance). However, we can also see results that show some concerning tendencies re garding their democratic assessment. Democratic competence of Estonian citizens is at 81%. 16 For most Estonians, it is important to live in a country that is governed democratically, but the rating 17 they ascribe to this importance is still the lowest among all seven countries. 84% of Estonians either“strongly agree” or“agree” that democracy is better than any other form of government, but as with many other statements, Estonians often refrain from giving strong ratings. Hence, they still rate democracy worse than respondents from all other countries under investigation(besides Ukrainians). The Estonian respondents also generally reject features of an authoritarian system: They are comparatively strongly opposed to“having the army rule”(the third highest average mean just behind Sweden and Germany). Yet, the statement “having a strong leader who does not have to bother with parliament or elections” is rejected by only 61% of all respondents. Only Ukrainians rate this statement more positively than Estonians. Similar to other countries, one of the more concerning findings is that 24.5% of Estonians regard it as rather or completely democratic 18 if the military removed a corrupt president. Furthermore, 58% of Estonian citizens find it“fairly good” or “very good” if“experts, not the government, made decisions for the country.” “How strongly do you agree that democracy is better than any other form of government?” All figures in% Sw e d en 49.5 42.5 6.0 G er ma ny 47.1 44.6 6.8 Spa in 44.4 44.7 8.8 P o la n d 32.7 56.5 9.2 S er b i a 24.4 60.9 11.9 E stoni a 23.5 60.8 13.7 Uk r a ine 15.1 37.9 38.6 8.4 20% 40% 60% 80% strongly agree agree disagree strongly disagree Due to rounding, totals may not correspond with the sum of the separate figures. Moreover, Estonians are not strongly concerned with the statement“The president encourages his supporters to disrupt opposition campaign rallies.” Less than 30% of the respondents rate this as“completely undemocratic” as opposed to over 40% in Poland, Germany, Sweden, Serbia and almost 40% in Spain. When it comes to religious freedom, Estonia ranks lowest: Estonian respondents rate the statement“Muslims are not allowed to publicly celebrate religious holidays” as more democratic than other countries do. 19 16 If we define a“democratically competent” respondent as someone who, on average, rates the four democratic items as more democratic than the ten undemocratic items present in the survey. 17 Reaching a mean score of 8.1 on a scale from 1 to 10. 18 Rating it with an 8, 9 or 10 on a scale from 1 to 10. 19 Reaching a mean score of 3.8 on a scale from 1(completely undemocratic) to 10(completely democratic). 40

IDENTITY, PARTISANSHIP, POLARIZATION “How good or bad do you find having a strong leader who does not have to bother with parliament or elections?” All figures in% Sw e d en G er ma n y Spa in Pola n d S er b i a E s to ni a 64.9 20.1 11.6 3.4 56.8 25.1 14.9 3.2 55.3 21.3 17.1 6.3 42.7 32.2 19.9 5.1 39.6 32.5 20.8 7.1 26.4 34.4 32.3 7.0 Uk r a ine 20.1 30.0 34.9 15.0 v er y bad 20% 40% 60% fa ir ly bad fa ir ly g ood 80% v er y g ood Due to rounding, totals may not correspond with the sum of the separate figures. “How democratic do you find it when a president encourages their supporters to disrupt the opposition's campaign rallies?” SRB SWE 50% GER POL 40% ESP UKR 30% EST 20% 10% 1 2 3 4 5 6 7 8 9 10 totally undemocratic totally democratic -7.8% In Estonia, the average electoral punishment for a candidate who adopts an undemocratic position is a loss of-7.8% of the overall vote share. WHAT KIND OF UNDEMOCRATIC BEHAVIOR DO ESTONIANS PUNISH? Estonian voters showed a willingness to punish undemocratic behavior. The average punishment is a loss of-7.8% of the overall vote share for the candidate who adopts an undemocratic position. The most punished undemocratic positions are“prosecuting journalists” with a -12.9% vote share loss and“violently disrupt opponent’s rallies” with a-11.4% vote share loss. On the other hand, Estonians punish the positions“passing laws without parliamentary debate” and“monitor social media” the least: the candidates who advocate for these undemocratic positions lose only-5.4% and-5.5% of their vote share respectively. Existence of trade-offs/compensation for undermining democracy punishment reward Average punishment -7.8% Partisan Same-sex couples’ rights Language Immigration Education Defense policy Environment EU-Russia -8% +23.6% +17.8% +13.3% 5.5% +11.5% 3.7% +8.6% 0.8% +7.9% 0.1% +7.4% -0.4% N/A -4% 0%+4%+8% 10.0% +12% Final reward 15.8% 41

How democratically elected politicians get away with autocratizing their country PARTY LOYALTY AND POLICY INTERESTS: WHICH ISSUES ARE MORE IMPORTANT THAN UPHOLDING DEMOCRATIC STANDARDS? Estonian voters forgive undemocratic behavior of a candidate when they are coming from their favorite party. They do the same when it comes to identity-based interests and defense policy, albeit to a significantly lower degree. Moreover, advocating for favorite socio-economic policies is only partly rewarded – and not much. Indeed, a candidate from the respondent’s favorite party who advocates for an undemocratic position is not punished, but instead ends up being rewarded with a+15.8% vote share increase. Favorite identity-based policies can compensate for undemocratic behavior as well. This is the case for all identity-related interests – same-sex couples’ rights, language and immigration – with a+10%,+5.5% and+3.7% final vote share increase respectively. The final rewarding rate with respect to advocating for the voter’s favorite policy on same-sex couples’ rights is the highest among all the seven countries under investigation. Contrary to partisan and identity-based interests, advocating for a voter’s favorite defense policy does not result in a clear reward as the final vote share of the undemocratic candidate increases by only 0.1%. Nevertheless, Estonia is the only country across the entire sample in which undemocratic behavior can be fully compensated by the proposition of a voter’s favorite defense policy. When it comes to socio-economic policy, the results are mixed. While an undemocratic candidate ends up being punished despite advocating for a voter’s favorite policy on environmental issue(-0.4% final vote share loss), they would be rewarded when advocating for a voter’s favorite policy regarding education(with a minor+0.8% vote share increase that is nevertheless the highest across the entire sample for this particular issue). ARE SUPPORTERS OF A CERTAIN POLICY INTEREST 20 MORE FORGIVING OF UNDEMOCRATIC BEHAVIOR THAN OTHERS? Same-sex couples’ rights Voters in Estonia reward politicians the most for their respective preferred position on same-sex couples’ rights. If we divide respondents along a left-leaning/ right-leaning axis, we can identify more nuanced findings about the willingness to punish undemocratic behavior. Concretely, when it comes to the issue of same-sex couples’ rights, respondents from both sides of the political spectrum show less attachment to democratic principles. However, those on the right showed the greatest willingness to forgive undemocratic behavior for it, with a final+13.3% vote share increase. The same tendency could be observed with moderate right voters on the issue of immigration(with a final+6.3% vote share increase). Interestingly, while left-leaning(+1.8% final vote share increase) and moderate left(+2.5% final vote share increase) voters in the end also reward an undemocratic candidate who advocates for their favorite immigration policy, voters on the right would neither reward, nor punish such a candidate. With respect to the language issue, those who favor the right to use the Russian language when communicating with state offices show more willingness to forgive undemocratic behavior, with a final+8.8% vote share increase. This is 4.6% more than in the case of those who prefer that only the Estonian language should be used. Moreover, voters who favor that Estonian defense policy is decided jointly by Estonia and the EU forgive undemocratic behavior, as in the end they reward such a candidate with a+5% vote share increase. The same can be said for those who prefer cutting personal income taxes and reducing education budget accordingly (+4.1% final vote share increase). 20 Please find an overview of all tested policy positions on pg. 13. 42

IDENTITY, PARTISANSHIP, POLARIZATION Defense Estonia is the only country in which voters forgive undemocratic behavior when a candidate proposes their favorite defense policy. Same-sex couples’ rights: shift in punishment when presented with favored policy punishment reward -13.3% +21.5% 8.2% Left -6.1% +11.0% 4.9% Center -12% -5.6% -8%-4% +18.9% 13.3% Right 0%+4%+8%+12% WHAT ROLE DOES POLARIZATION PLAY? When it comes to same-sex couples’ rights, we can see that there is a correlation between the high polarization on this issue and the willingness to forgive undemocratic behavior for it. As a matter of fact, the issue of rights for same-sex couples is the most polarizing issue in Estonia. Moreover, we can observe the same correlation regarding the is sue of immigration. At the same time, such a correlation cannot be observed for the majority of issues for which the Estonians would forgive undemocratic behavior, as language, education and defense policy seem not to be polarizing issues in the Estonian society. Nevertheless, voters also forgive undemocratic behavior for them. WHO ARE THE MAIN‘DEFENDERS OF DEMOCRACY’ IN ESTONIA? Average electoral punishment by respondents’ age groups 18-29 years 30-39 40-49 50-65 66-75 -2% -4% -6% -8% -10% -12% Estonia All countries* average -14% *No data was collected for age group 66-75 in Serbia and Ukraine. The main defenders of democracy in Estonia are – similar to most other cases – young people, students and women. The age distribution is very linear, the younger voters are, the more they punish. The gender difference is somewhat more pronounced than the average across all countries. Other defenders of democracy include people who have a job(part or full time) and who take care of their family. In turn, the willingness to punish politicians who violate democratic principles is significantly lower among the unemployed and those with a more negative economic assessment of their family. It is also not surprising that those who are more interested in politics, who are democratically competent, who are more trustful, who are less religious and who lack an authoritarian personality also show a greater willingness to punish undemocratic behavior. 43

How democratically elected politicians get away with autocratizing their country WHICH PARTY‘S VOTERS PUNISH UNDEMOCRATIC BEHAVIOR THE MOST? When it comes to party affiliation, voters of all Estonian parties show a general willingness to punish undemocratic politicians, however, one can observe significant variations: voters of the far-right EKRE and voters of the ruling party Centre punish the least, with a-4.1% and-4.5% vote share loss respectively. On the other hand, voters of the SDE, Isamaa and the other ruling party Reform punish much more, as the vote share loss is-13.2%,-9.4%, and-10.8% respectively. Interestingly, the punishment rate among the respondents who would not vote for either of these five parties is the second highest, with-12%. Average punishment for undermining democracy punishment SDE Reform Isamaa -13.2% -10.8% -9.4% Centre -4.5% EKRE -2% -4% -4.1% -6% -8% -10% -12% Summary for Estonia A majority of respondents in Estonia support democratic standards and the rule of law. However, compared with the other countries in our study, Estonians rate some undemocratic positions as significantly more democratic. In our experiment, however, we see that Estonians punish a candidate’s undemocratic behavior with-7.8% loss of votes, slightly below the average of the cross-country-sample. They are most likely to defend the right of assembly and electoral fairness as well as the freedom of the press. As in all other countries under investigation, party loyalty is the most important determinant of voting behavior. Estonians completely forgive a candidate from their favorite party their undemocratic behavior. But group loyalty and identity issues also play an important role when it comes to voting decisions in Estonia: Issues such as rights for same-sex couples, language as well as immigration are prioritized at the ballot box over rejecting undemocratic positions. The willingness to forgive undemocratic behavior for particular interests, however, only partly correlates with the degree of polarization for these issues. Indeed, the majority of issues for which the Estonians forgive undemocratic behavior seem not to be polarizing in the Estonian society. Estonian voters however do regard all these issues as important, which could be one of the explanations for the willingness to disregard democratic principles. 44

IDENTITY, PARTISANSHIP, POLARIZATION Facts& Figures Population Political system Varieties of Democracy regime type Which issues let politicians get away with undemocratic behavior? Issues displayed according 8 to their importance to voters and degree of polarization ! Issues for which voters 7 forgive undemocratic behavior 6 ! Education ! Environment Immigration ! Same-sex ! couples’ rights Defense policy 5 more important (1 10) Varieties of Democracy Liberal Democracy Index ranking(2021) 4 3 0.2 0.3 0.4 0.5 0.6 more polarizing (0.0- 1.0) 82% Democratic competence of German respondents Percentage of respondents who on average rate the four democratic items as more democratic than the ten undemocratic items in our survey How much are key elements of liberal democracy valued? The higher the vote loss for an undemocratic policy proposed by a political candidate, the more a certain democratic element is valued. -4.4% -3.3% -6.1% -6.5% -6.3% -7.1% -11.1% -8.4% -12.4% Ban foreign union funding Pass laws without parliamentary debate Ban foreign NGO funding Monitor social media Partisan infrastructure spending Discipline judges Ban prayers for Muslims Prosecute journalists Violently disrupt rallies -20%-15%-10%-5% 0% Average punishment across all 7 countries Min/Max punishment across all 7 countries Punishment in Germany 45

German parties presented in the survey How democratically elected politicians get away with autocratizing their country English name German name Social Democratic Party of Germany Sozialdemokratische Partei Deutschlands(SPD) Alliance 90/ The Greens Bündnis 90/ Die Grünen (GRÜNE) CDU-CSU Democratic Party Democratic Union Free Christian Christian Social of Germany Union in Bavaria Alternative for Germany Freie Demokratische Partei(FDP) Christlich Demokratische Union Deutschlands (CDU) Christlich-Soziale Union in Bayern (CSU) Alternative für Deutschland(AfD) affiliation European S&D Greens/EFA ALDE EPP EPP ID Currently governing The Left Die Linke GUE/NGL KNOWLEDGE ABOUT AND SUPPORT FOR DEMOCRACY Across the entire sample, Germans have a very good understanding of what democracy is and show a very high support for democracy, ranking just behind Sweden in both aspects. 91.7% of Germans“agree” or“strongly agree” that“Democracy may have problems, but it is better than any other form of government.” More than half of all German respondents(56.8%) rate the statement“Having a strong leader who does not have to bother with parliament or elections” as“very bad.” Still, almost one-fifth of Germans think such a situation is either“fairly good”(14.9%) or“very good”(3.2%). A slight majority of Germans (38.4%) finds it“rather good” to have“experts, not the government, make decisions for the country.” The recent experience of the pandemic may have led people to accept to be governed by experts rather than an elected government. Germans and Swedes evaluate their countries as being the most democratic. However, on a scale from 1 to 10 both countries’ citizens were reluctant to give their countries absolute ratings: this resulted in a means of 6.5 for Germany and 6.8 for Sweden. When respondents in Germany were asked to assess how democratic they find certain actions(democratic competence test), there were a few findings that stood out: Compared to the other countries under investigation, Germans trust their high court the most: the statement “How democratic do you find the practice of the country’s judiciary being staffed with individuals loyal to the governing party?” SRB POL 40% ESP SWE GER 30% EST UKR 20% 10% 1 2 3 4 5 6 7 8 9 10 totally undemocratic totally democratic “How democratic do you find the practice of journalists frequently disagreeing with the president's policies?” POL 25% SWE SRB 20% GER UKR 15% ESP EST 10% 5% 1 2 3 4 5 6 7 8 9 10 totally undemocratic totally democratic -7.5% This is the average punishment corresponding to the share of German voters willing to defect from an otherwise favored candidate once they adopt an undemocratic position. 46

IDENTITY, PARTISANSHIP, POLARIZATION “The country’s high court ruled that a government policy was unconstitutional” was ranked as more democratic than in other countries. On a scale from 1(not at all democratic) to 10(completely democratic), Germans gave this statement on average a 6.6. At the same time, Germans are comparatively uncritical if the judiciary were staffed with judges loyal to the governing party: similar to Ukrainians and Estonians, only about 30% of all German respondents consider this practice to be“not at all democratic.” Germans are also the ones who most strongly agreed that losing an election can be an inherent part of a democracy, with the statement“The prime minister conceded a narrow elec tion” receiving a mean score of 7.0. Journalists who frequently disagree with the president’s policies, however, are comparatively less appreciated in Germany than in other countries(6.4). If journalists were prosecuted for criticizing the government though, Germans view this as more undemocratic than respondents in most of the other countries(1.9). The most concerning finding is that 26% of Germans regard it as rather or completely democratic 21 if the military removed a corrupt president. With a mean score of 4.6, this statement is certainly viewed as less democratic than some other statements, but Germans also assess a military coup on a corrupt president more democratic than having unelected experts govern the country(2.9) or having the government cut spending in regions that voted for the opposition party(2.4). These results can of course also be an expression of how much Germans(similar to their European peers) dislike corruption. WHAT KIND OF UNDEMOCRATIC BEHAVIOR DO GERMANS PUNISH? Germans are willing to punish candidates with undemocratic positions, as is confirmed by an average of a-7.5% loss for these candidates. This rate of vote loss is just below the average rate of punishment for the entire country sample(-7.8%) The issues that Germans find worth punishing most are a candidate’s encouragement of their supporters to violently disrupt campaign rallies of their political opponents(resulting in loss of vote share of-12.4%) and a candidate saying that Muslims should not be allowed to pray during their breaks at work (-11.1%). The undemocratic statement that Germans punish the least at the ballot box is the suggestion to pass laws without parliamentary debate if criticized by the opposition(only -3.3%). Interestingly, other established Western democracies had a similar low punishing rate on this issue(Sweden-4.9 and Spain-2.5%). In Poland, on the other hand, this statement is punished with a-14.5% loss. Germans also do not punish much if a candidate wanted to ban foreign funding for labor unions(-4.4%) or foreign non-governmental organizations (-6.1%). Existence of trade-offs/compensation for undermining democracy punishment reward Average punishment -7.5% Partisan Same-sex couples’ rights Environment Immigration Education Defense policy Language EU-Russia -8% +22.9% +13.3% 5.8% +11.1% 3.6% +10.2% 2.7% +8.1% 0.6% +7.0% -0.5% 7.4% N/A N/A -4% 0%+4%+8%+12% Final reward 15.4% 21 Rating it with an 8, 9 or 10 on a scale from 1 to 10. 47

How democratically elected politicians get away with autocratizing their country PARTY LOYALTY AND POLICY INTERESTS: WHICH ISSUES ARE MORE IMPORTANT THAN UPHOLDING DEMOCRATIC STANDARDS? First and foremost, Germans accept a possible violation of democratic norms if that means their favorite party wins: the average loss in votes of-7.5% of a candidate who proposes an undemocratic action is compensated with an increase of +22.9% in votes if the candidate belongs to the party that the voter supports. All in all, this situation thus still results in a +15.4% win. Identity-based and socio-economic policy positions play a significantly less important role, but the voter’s favorite position still compensates entirely for the undemocratic position, resulting in a+5.8% win for the voter’s favorite position on same-sex couples’ rights, and in a+2.7% win for the voter’s favorite immigration policy. The voter’s favorite tax policy in combination with the voter’s environmental policy leads to a+3.6% win – and in combination with the voter’s favorite position on education spending to a win of+0.6%. The favorite defense policy cannot entirely compensate for an undemocratic position, as it results in a-0.5% vote loss. ARE SUPPORTERS OF A CERTAIN POLICY INTEREST 22 MORE FORGIVING OF UNDEMOCRATIC BEHAVIOR THAN OTHERS? Voters with a left-leaning position when it comes to economic issues punish more(-8.9%) than voters with a right-leaning position(-3.4%). When presented with their respective favorite position on taxes and education, we see that both camps become more tolerating of an undemocratic candidate: left-leaning voters even reward the undemocratic candidate with+0.4%, while the right-leaning voters reward them with +6.7% respectively, thus deviating slightly more than the left-leaning voters. Green voters generally punish undemocratic behavior almost twice as much as conservative voters(-9.7% vs.-5.9%). But when presented with their favorite policy on renewable energies, both green and conservative voters in Germany reward that policy and forgive undemocratic behavior(+0.7% for green voters vs.+5.9% for conservative voters). If we have a closer look at the voters’ favorite position on sexual minorities we find the following interesting fact for Germany: Voters who are in favor of extending more rights to same-sex couples generally punish a candidate above average, with a loss of-9.6%, whereas people who have more conservative views punish below average, namely-4.5%(i.e., voters who are pro equal marriage rights but against adoption rights) and-3.6% respectively(i.e. voters who are against both marriage and adoption rights). When voters are confronted with a candidate that holds their favorite position on same-sex couples’ rights and an undemocratic position at the same time, then the liberal voters punish much less(+7.1% vote share win) compared to voters who are only in favor of marriage equality and voters who are against any equal rights for same-sex couples(1.8% and+5.0% win respectively). In general, left-leaning voters on the issue of immigration punish undemocratic behavior more with a-11.2% vote share loss; right-leaning voters only punish it with-3.1%, and thus also less than average. When voters are however presented with their favorite policy on immigration, those left-leaning suddenly reward the candidate with+3.0%, whereas rightwing voters end up punishing by just as much as center-left voters,-1.3% and-1.4% respectively. However, voters with a center-right position still reward the most when presented with their favorite position on immigration, namely with+4.8%. None of the 2500 survey respondents chose the right-leaning position that“Defense policy for Germany should be decided by German institutions alone.” Respondents with the view that defense policy should be a shared task between Germany and the EU generally tolerate undemocratic behavior less than respondents who prefer sole EU responsibility for defense (-6.4% punishment rate vs.-3.4%). However, when presented with their preferred policy on defense policy, respondents with a cooperative view become significantly more forgiving of undemocratic behavior(rewarding it with+4.4%) than respondents who want to see defense policy in the hands of the EU alone(-1.8%). This could indicate that defense policy is a less important topic for the pure supra-nationalists. 22 Please find an overview of all tested policy positions on pg. 13. 48

IDENTITY, PARTISANSHIP, POLARIZATION Average electoral punishment by respondents’ age groups 18-29 years 30-39 40-49 50-65 66-75 -2% -4% -6% -8% -10% -12% Germany All countries* average -14% *No data was collected for age group 66-75 in Serbia and Ukraine. Immigration: shift in punishment when presented with favored policy punishment reward -11.2% -9.0% +14.2% +7.6% -1.4% 3.0% Left CenterLeft -5.9% +10.7% +1.8% -1.3% -3.1% 4.8% CenterRight Right -8%-4% 0%+4% WHAT ROLE DOES POLARIZATION PLAY? In Germany, the most polarizing issue is how to tax renewable energies. As we have seen above, this is also an issue that voters would disregard undemocratic behavior for. Moreover, we find a similar correlation for the policy on same-sex couples’ rights. Although it is slightly less polarizing than the environmental issue, Germans are readier to tolerate undemocratic behavior for it. Hence, for these two issues we can observe a correlation between an increased polarization and the willingness to forgive undemocratic behavior. Interestingly, we found that the issue of immigration, which also leads to a greater tolerance of undemocratic behavior, is not that polarizing in Germany. The same holds for issue revolving around education. Nonpartisans Germans who do not favor any of the six parliamentary parties showed the greatest willingness to punish undemocratic behavior. WHO ARE THE MAIN‘DEFENDERS OF DEMOCRACY’ IN GERMANY? As in most of the other countries under investigation, women punish undemocratic behavior more than men. The same holds true for those who showed more interest in politics, those who are democratically competent, those who are less religious, those who are more trusting of other people and those who lack an authoritarian personality. When it comes to education, the more educated punish slightly more, but just like in Spain and Sweden, the level of education does not have a great effect on the degree of punishing undemocratic behavior, especially not when comparing the results for voters with a medium and high level of education. The youngest voter group(18-29 years of age) in Germany punishes undemocratic behavior distinctly less than the same age group in all other EU countries under investigation. Germany’s 50 to 65-year-olds however are the strongest group when it comes to defending democracy, both within Germany but also when compared to that age-group’s average among all countries under investigation. Also, the voters’ economic situation does have a slight influence on how strongly people defend democracy: Voters who assess their country’s economic situation better punish undemocratic behavior more than Germans with a worse assessment on Germany’s economy and also more than people with a similarly positive assessment of their country’s economy in all other countries. The same correlation can also be found for voters’ behavior and the assessment of their own family’s economic situation in Germany. Like in other countries, students punish undemocratic behavior the most, but in Germany full-time employed and unemployed people punish to an equal degree, while part-time employees and the retired punish slightly more. 49

How democratically elected politicians get away with autocratizing their country WHICH PARTY‘S VOTERS PUNISH UNDEMOCRATIC BEHAVIOR THE MOST? Voters of all parties represented in the German national parliament punish undemocratic behavior of a candidate to a certain extent, however, the punishing rates among the parties differ quite a bit. A candidate with an undemocratic position would lose-9.7% of votes from Green voters and-8.6% from Left party voters. Among voters of the FDP and SPD, that candidate loses -7% and-7.6% respectively, while AfD voters punish the candidate with a-5.3% loss and CDU/CSU voters with a-5.2% loss. The latter voter group also rewards passing laws without parliamentary debate with a+12.1% vote share increase. Interestingly, voters who vote for “other” parties punish the strongest:-11.1% of the vote share. And even non-voters punish with -4.4% of the vote share. Average punishment for undermining democracy Bündnis 90/ Die Grünen Die Linke SPD FDP AfD CDU/CSU -2% Environment & taxes German voters reward the highest across the sample when a candidate proposes their respective favorite policy on the environment and taxes, thereby neglecting undemocratic behavior. punishment -9.7% -8.6% -7.6% -7.0% -5.3% -4% -5.2% -6% -8% Summary for Germany A large majority of respondents in Germany support democratic standards and the rule of law. However, we also see some ambivalences in our findings: Germans are comparatively trustful of the judicial system, however, also find it comparatively acceptable if judges are appointed according to partisan interests. When it comes to defending democratic standards in our experiment at the ballot box, Germans score slightly below average. They are most likely to defend religious freedom, the right of assembly and electoral fairness. As for all other countries under observation, however, party loyalty is the most important determinant of voting behavior. Germans completely forgive a candidate their undemocratic behavior if they are from their favorite party. But group loyalty and identity issues also play an important role when it comes to voting decisions in Germany: a preferred position on rights for same-sex couples as well as on immigration are prioritized at the ballot box over rejecting undemocratic positions. Polarization for certain issues appears to play only a partial role in punishing undemocratic behavior. In Germany, there are some issues that are not polarizing but are rated as important enough to compensate for the loss of votes for undemocratic behavior. 50

IDENTITY, PARTISANSHIP, POLARIZATION Facts& Figures Population Political system Varieties of Democracy regime type Which issues let politicians get away with undemocratic behavior? Issues displayed according 8 to their importance to voters and degree of polarization ! Issues for which voters 7 forgive undemocratic behavior Defense policy Education Environment 6 Immigration Same-sex ! 5 couples’ rights more important (1 10) Varieties of Democracy Liberal Democracy Index ranking(2021) 4 3 0.2 0.3 0.4 0.5 0.6 more polarizing (0.0- 1.0) 78% Democratic competence of Polish respondents Percentage of respondents who on average rate the four democratic items as more democratic than the ten undemocratic items in our survey How much are key elements of liberal democracy valued? The higher the vote loss for an undemocratic policy proposed by a political candidate, the more a certain democratic element is valued. -14.5% -18.2% -14.1% -11.2% -4.6% -7.6% -7.6% -4.1% -7.4% Ban foreign union funding Pass laws without parliamentary debate Ban foreign NGO funding Monitor social media Partisan infrastructure spending Discipline judges Ban prayers for Muslims Prosecute journalists Violently disrupt rallies -20%-15%-10%-5% 0% Average punishment across all 7 countries Min/Max punishment across all 7 countries Punishment in Poland 51

How democratically elected politicians get away with autocratizing their country Polish parties presented in the survey English name Polish name European affiliation Currently governing Law and Justice Civic Platform Prawo i Sprawiedliwość(PiS) Platforma Obywatelska (PO) ECR EPP Poland 2050 Polska 2050 ALDE Confederation Konfederacja ID The Left Lewica S&D KNOWLEDGE ABOUT AND SUPPORT FOR DEMOCRACY On average, Polish respondents have a very good understanding of what democracy is and show a fairly high support for democracy; yet, there are some surprising findings when we have a closer look. 89.2% of Poles“agree” or“strongly agree” that“Democracy may have problems, but it is better than any other form of government.” This means that Poles are certainly less convinced of democracy than Germans, Swedes and Spaniards. Moreover, 42.7% of Poles rate the statement“Having a strong leader who does not have to bother with parliament or elections” as“very bad” and 32.2% still rate it as“fairly bad.” A clear majority of Poles(45.6%) find it“rather good” to have“experts, not the government, make decisions for the country.” On the one hand, as is the case in other countries, the pandemic experience may have normalized the idea of being governed by experts among the Polish population, rather than by an elected government. On the other hand, this could be interpreted as a lack of confidence in the political elite, as Poles are the second most unsatisfied with democracy in their own country – only the Serbs are more unsatisfied. It is also noteworthy that a majority of Poles think that their country is not being governed democratically at all, with 41.8% giving it a rating between 1 to 3 23 on a scale from 1 to 10. And 44.8% of the people are“not satisfied” or“not at all satisfied” with the way democracy works in Poland. Compared to the other countries under investigation, Poles do trust their high court a lot(only Germans are more trusting), viewing the statement“The country’s high court ruled that a government policy was unconstitutional” as more democratic than other citizens do. This finding is remarkable, particularly in light of the recent politicization of courts in Poland. 24 Moreover, more Poles than respondents in all other countries(26.4%) rate journalists who frequently disagree with the president’s policies as completely democratic. Additionally, Poles are the most sympathetic to the democratic practice of the opposition organising protests against the government, with almost 28.8% of Poles giving it the highest rating on the 1-10 scale. “How democratically is your country being governed today?” 20% 15% SWE GER 10% POL SRB ESP 5% UKR EST 1 2 3 4 5 6 7 8 9 10 totally undemocratic totally democratic “How democratic do you find the practice of opposition parties organizing protests against the government?” POL 25% SRB SWE 20% GER ESP 15% UKR EST 10% 5% 1 2 3 4 5 6 7 8 9 10 totally undemocratic totally democratic 23 1(meaning“not at all democratic”) was chosen by 17.07%, a rating of 2 by 13.08% and a rating of 3 by 11.65% of all the people asked. 24 Giving it a rating between 1 and 3 on the scale from 1 to 10. 52

IDENTITY, PARTISANSHIP, POLARIZATION The most concerning finding is that more than 33.3% of Poles regard it as rather or completely democratic 25 if the military removed a corrupt president. This is the highest support among all countries under investigation. Also, more than 40% of Poles find it“not at all democratic” to pass a law in parliament without debate. This percentage is only surpassed by the Serbs. And Polish respondents hold the same very critical view when rating the practice of staffing the country’s judiciary with judges loyal to the governing party: almost 40% of Poles rate this scenario with a 1(on the scale from 1 to 10), which is in fact more than in Germany or even Sweden. It is also noteworthy that far more Poles than any of the other countries’ citizens find it“not at all democratic” if the“government banned civil society organizations that receive funds from abroad from operating inside the country”: over 35.6% of the Polish people chose this rating compared to just over 20% in Germany or just over 25% in Sweden. Poles are also significantly more critical if the“country’s security agencies collect data on their citizens’ internet activity”: Over 33% of the Polish survey responders rated this practice with a 1, a percentage that is only exceeded by the Serbs(40.4%), while Germans, Spaniards or Estonians scored significantly lower(with around 25%). WHAT KIND OF UNDEMOCRATIC BEHAVIOR DO POLES PUNISH? Poles are willing to punish candidates with undemocratic positions, as such behavior results in an average of a-9.2% loss of votes for these candidates. This rate of vote loss is the second highest just behind Sweden and is also above the average rate of punishment of the entire country sample(-7.8%). At first glance, this comparatively high punishing rate of the Polish respondents seems surprising, especially when considering that the currently ruling government of the Law and Justice(PiS) party has undertaken several steps that have weakened the country’s democracy and rule of law(justice reform, restrictions on the media and civil society). The reaction to the government’s restrictions from within the country(by politicians of the opposition, civil society, the independent media etc.) as well as from the EU may actually have raised more awareness of undemocratic practices than in other European societies. The issues that Poles find worth punishing most are issues of civil rights and checks and balances: A candidate saying that the government should prosecute journalists who accuse the prime minister of corruption is punished with a vote share loss of-18.2%, while a candidate who thinks that“the government should discipline judges who publicly criticize it” loses-14.1%. These are the highest percentages among all countries under observation. Poles also punish a candidate who suggests to “pass laws without parliamentary debate if criticized by the opposition” with a vote loss of-14.5%. Interestingly, established Western democracies have a comparatively low punishing rate for this issue(Sweden-4.9%, Germany-3.3% and Spain -2.5%). Like respondents in other countries, Poles do not punish much if a candidate wants to ban foreign funding for labour unions(-4.6%) or foreign non-governmental organizations (-4.1%). Existence of trade-offs/compensation for undermining democracy punishment reward Average punishment -9.2% Partisan +29.2% Same-sex couples’ rights +11.2% 2.0% Environment +8.3% -0.9% Immigration +7.5% -1.7% Defense policy +6.8% -2.4% Education +5.8% -3.4% Language N/A EU-Russia N/A -8%-4% 0%+4%+8%+12%+16% Final reward 20.0% 25 Rating it with an 8, 9 or 10 on a scale from 1 to 10. 53

How democratically elected politicians get away with autocratizing their country PARTY LOYALTY AND POLICY INTERESTS: WHICH ISSUES ARE MORE IMPORTANT THAN UPHOLDING DEMOCRATIC STANDARDS? First and foremost, Poles accept a violation of democratic norms if this means their favorite party wins: the average loss in votes of-9.2% of a candidate who proposes an undemocratic ac tion is compensated with a+29.2% in vote boost if the candidate belongs to the party that the voter supports, which means the candidate still ends up with a+20% win in vote shares. All other interests play a significantly lower role. Only the voters’ favorite position on an identity policy compensates entirely for the undemocratic position, resulting in a+2% win for the voters’ favorite position on marriage and adoption rights for same-sex couples. All other policies cannot entirely compensate: The voter’s favorite tax policy in combination with a favored environmental policy leads to a-0.9% loss and in combination with a favored position on education spending to a loss of-3.4%. The favorite immigration policy still results in a-1.7% vote loss, and the favorite defense policy results in a -2.4% vote loss for the undemocratic position. Environment Only in Poland an issue that is both polarizing and salient could be identified for which voters do not forgive undemocratic behavior: environmental policy. ARE SUPPORTERS OF A CERTAIN POLICY INTEREST 26 MORE FORGIVING OF UNDEMOCRATIC BEHAVIOR THAN OTHERS? If we have a closer look at the voters’ preferred policy position on sexual minorities we find the following interesting correlation for Poland: Voters who are in favor of an extension of rights for sexual minorities punish a candidate generally above average with a loss of-13.0%, while voters who support equal marriage rights but are against adoption rights still punish slightly above average(-10.0%). Conservative voters(who are against both marriage and adoption rights) punish less than half of the average, namely with a loss of-4.3%. When voters are confronted with a candidate that holds their favorite position on sexual minorities’ rights and an undemocratic position at the same time, however, the liberal voters punished less, resulting in a mere-2.0% vote share loss, compared with a+1.4% win for the voters in favor of only marriage equality and a +8.5% win for the voters against any equal rights for same-sex couples. In general, left-leaning voters on immigration punish undemocratic behavior at the ballot box with a-11.6% loss of the vote share as opposed to right-leaning voters who punish it with -5.8%, less than average. When presented with their favorite policy on immigration, however, left-leaning respondents punish significantly less than before:-2.7%, whereas right-wing voters end up punishing by just as much as Same-sex couples’ rights: shift in punishment when presented with favored policy punishment reward -13.0% +11.0% -2.0% Left -10.0% +11.4% 1.4% Center -12%-8% -4.3% -4% +12.8% 0%+4% 8.5% Right +8% mid-left voters,-0.7%. Only 40 of the 1500 survey respondents chose the right-leaning position that declared that“Defense policy for Poland should be decided by Polish institutions alone.” Respondents with the view that defense policy should be a common task of Poland and the EU together generally tolerate undemocratic behavior more than respondents who prefer sole EU responsibility(-7.6% punishment rate vs.-9.4%). And when presented with their preferred defense policy, respondents with a cooperative view bec o m e s i g n ifi c a n t l y m o re f o r g i v i n g o f undemocratic behavior(rewarding it with +2.1%) than respondents who want to see deImmigration Only in Poland voters do not forgive undemocratic behavior for immigration-related interests. 26 Please find an overview of all tested policy positions on pg. 13. 54

IDENTITY, PARTISANSHIP, POLARIZATION fense policy in the hands of the EU alone(-6.0%). This could indicate that defense policy is a less important topic for supranationalists. Respondents with a libertarian view are the most forgiving of undemocratic behavior of candidates in our survey, but they nevertheless vote less for undemocratic candidates(-6.5%). Undemocratic candidates lose-9.9% of the votes from people who want to keep the income tax rates and the education budget the way they are; and the same candidates lose-10.9% of their vote share of people with a more left-wing economic position. When presented with their favorite policy on taxes and education, however, all groups become very forgiving of undemocratic behavior, but with slight differences: left-leaning people still make the candidates lose-6.5% of the votes, people with a position in the middle-4.9% of the vote, and people with a libertarian position do not make the candidates lose any votes(0.0%). Green voters generally punish undemocratic behavior more than conservative voters(-12.6% vs.-7.7%). But when presented with their respective favorite policy on renewable energies, conservative voters become twice as tolerant towards undemocratic candidates than green voters: They reward the candidate with+3.1%, while green voters still punish by-7.3%. WHAT ROLE DOES POLARIZATION PLAY? -9.2% This is the average punishment corresponding to the share of Polish voters willing to defect from an otherwise favored candidate once they adopt an undemocratic position. Average electoral punishment by assessment of the country’s economic situation 16% 12% 8% 4% 0% -8% -10% -12% Poland All countries average 23 bad 45678 assessed economic situation 9 10 good We find that in Poland the most polarizing issue is which rights to grant to same-sex couples. As we have seen above, this is also an issue that voters disregard undemocratic behavior for. Hence, for this particular issue we can observe a correlation between an increased polarization and the willingness to forgive undemocratic behavior. Moreover, only in Poland we can identify a polarizing and at the same time salient issue for which, however, voters do not forgive undemocratic behavior: renewable vs. fossil energies. WHO ARE THE MAIN‘DEFENDERS OF DEMOCRACY’ IN POLAND? As in most of the other countries under investigation, women, students, and the more educated punish undemocratic behavior more than comparative sub-group members. The same holds true for those who are more interested in politics and have more political knowledge, those who are democratically competent, those who are less religious and those who lack an authoritarian personality. Notably, whether a person is trustworthy of other people does not influence the punishment rate of undemocratic behavior in any way. Just like in Spain, Estonia and Sweden, the youngest voter group(18-29 ys.) defends democracy more strongly than young people in Germany, Serbia and Ukraine. Poland moreover stands out from all other countries in another interesting way: Voters who assess their country’s economic situation worse punish un democratic behavior significantly more than people with a more positive view with regards to Poland’s economy. The former also punish by far more than citizens of all other countries who hold a similarly bleak assessment of their country’s economy. And the same correlation can be found for voting behavior and the assessment of the economic situation of one’s family in Poland. And who shows less willingness to punish politicians who violate democratic principles? Those with a more positive economic assessment of the country and their family. 55

How democratically elected politicians get away with autocratizing their country WHICH PARTY‘S VOTERS PUNISH UNDEMOCRATIC BEHAVIOR THE MOST? Voters of all Polish parties punish undemocratic behavior, yet to a very different extent: A candidate with an undemocratic position loses-13.5% from voters affiliated with the social democratic Lewica, closely followed by-13.3% from voters of the conservative Civic Platform(PO) and-13% from voters of the centrist Poland2050. Among voters of the far-right Konfederacja party, the very same candidate loses-5.5%, while among voters of the other far-right party(PiS) they lose even less,-3.4%. Interestingly, the latter voters also reward several undemocratic positions:“to ban foreign NGO funding” with+4.3%,“to prosecute journalists”with+20.8%,“to pass laws without parliamentary debate” with +26.2% and“to discipline judges who criticize the government” with a+29% vote share increase. The fact that attacks on judicial independence enjoy such popularity among PiS voters sheds new light on the reasons for the successful undermining of the rule of law in Poland. Voters who vote for“other” parties punish with a loss of-10.8% of the vote share, whereas non-voters punish with-5.4%. Average punishment for undermining democracy PiS PiS voters showed the least willingness to punish undemocratic behavior across the entire sample. punishment Civic Platform(PO) Lewica -13.5% -13.3% Polska2050 Konfederacja -5.5% -13.1% PiS -2% -3.4% -4% -6% -8% -10% -12% Summary for Poland A large majority of Poles value the various facets of democracy comparatively highly. However, they are quite critical of the democratic system, and especially of the way democracy functions in their own country. Compared to the other countries under investigation, they regard the judicial system and freedom of the media as highly important. Poles express their high regard for democratic standards and the rule of law not only in their responses to the direct questions, but also in the candidate choice experiments: they show more strongly than most other countries(except Sweden) that they would punish undemocratic behavior at the ballot box. Poles are most likely to defend checks and balances and the freedom of the press. As in all other countries under investigation, however, party loyalty is an important determinant of voting behavior. Poles completely forgive a candidate their undemocratic behavior if they are from their favorite party. In some respects, also identity issues and polarization influence voting decisions: regarding the question of rights for same-sex couples, a voter’s preferred position on this issue is placed above the motivation to punish undemocratic behavior. In Poland we also found an issue(environmental protection) that, while polarizing and salient, is nevertheless not a decisive factor when it comes to forgiving undemocratic behavior. 56

IDENTITY, PARTISANSHIP, POLARIZATION Facts& Figures Population Political system Which issues let politicians get away with undemocratic behavior? Issues displayed according 8 to their importance to voters and degree of polarization ! Issues for which voters 7 forgive undemocratic behavior EU-Russia ! ! Environment Education ! Language 6 Varieties of Democracy regime type 5 more important (1 10) Varieties of Democracy Liberal Democracy Index ranking(2021) 4 3 0.2 Same-sex ! couples’ rights 0.3 0.4 0.5 0.6 more polarizing (0.0- 1.0) 80% Democratic competence of Serbian respondents Percentage of respondents who on average rate the four democratic items as more democratic than the ten undemocratic items in our survey How much are key elements of liberal democracy valued? The higher the vote loss for an undemocratic policy proposed by a political candidate, the more a certain democratic element is valued. -4.8% -9.7% -6.3% -5.1% -5.1% -8.1% -12.6% -6.3% -11.8% Ban foreign union funding Pass laws without parliamentary debate Ban foreign NGO funding Monitor social media Partisan infrastructure spending Discipline judges Ban prayers for Muslims Prosecute journalists Violently disrupt rallies -20%-15%-10%-5% 0% Average punishment across all 7 countries Min/Max punishment across all 7 countries Punishment in Serbia 57

How democratically elected politicians get away with autocratizing their country Serbian parties presented in the survey English name Serbian name Serbian Progressive Party Socialist Party of Democratic Party Serbia Srpska Napredna Stranka(SNS) Socijalistička Partija Srbije(SPS) Demokratska Stranka(DS) European affiliation EPP(associated) S&D(associated) Currently governing Dveri Dveri Party of Freedom and Justice Stranka Slobode i Pravde(SSP) It is enough Dosta je Bilo ECR KNOWLEDGE ABOUT AND SUPPORT FOR DEMOCRACY Respondents of Serbia have a good understanding of what democracy is and what it is not. Indeed, democratic competence of Serbian citizens is at 80%. They are pro-democracy oriented as well, but at the same time unsatisfied with how Serbian democracy performs. For the Serbs, it is important to live in a country that is governed democratically(mean score 8.2), to have a democratic political system(78% rank this as either“fairly good” or“very good”). Moreover, 85% of all Serb respondents either“strongly agree” or“agree” that democracy is better than any other form of government. Interestingly, Serbs seem to appreciate democracy generally slightly more than the respondents in the EU-member state Estonia. The Serbian respondents also generally reject features of an authoritarian system: They are opposed to“having the army rule”(80% ranked this as either“bad” or“fairly bad”). And the statement“having a strong leader who does not have to bother with parliament or elections” is rated negatively by 72% – that is 11% more than in Estonia. One of the more concerning findings is that 29.2% of Serbs regard it as rather or completely democratic 27 if the military removed a corrupt president. We moreover found that Serbian respondents appreciate especially the“electoral fairness” aspect of democracy: One undemocratic practice that is very much disliked by Serbs is “passing laws without a debate in parliament.” This item received the highest“completely undemocratic” rating in contrast to the other seven countries under investigation. Serbs also agree that losing an election can be an inherent part of a democracy: the statement“The prime minister conceded a nar“How good or bad do you find having a strong leader who does not have to bother with parliament or elections?” All figures in% Sw e d en G er ma n y Spa in Pola n d S er b i a E s to ni a 64.9 20.1 11.6 3.4 56.8 25.1 14.9 3.2 55.3 21.3 17.1 6.3 42.7 32.2 19.9 5.1 39.6 32.5 20.8 7.1 26.4 34.4 32.3 7.0 Uk r a ine 20.1 30.0 34.9 15.0 v er y bad 20% 40% 60% fa ir ly bad fa ir ly g ood 80% v er y g ood Due to rounding, totals may not correspond with the sum of the separate figures. row election” got the second highest mean score of all European countries(6.8). And Serbs also rate it as very democratic if opposition parties organize protests against the governing party, by giving it the same mean score as the Swedes: 6.4. Serbs also view it as rather democratic if the country’s high court ruled that a government policy was unconstitutional, ranking this statement with the second highest mean score (6.0) just behind Germany. However, it appears that the Serbian democracy is currently not fulfilling the expectations of the Serbian citizens. Across the entire sample, the Serbs see their country as being governed 27 Rating it with an 8, 9 or 10 on a scale from 1 to 10. 58

IDENTITY, PARTISANSHIP, POLARIZATION Foreign policy The dispute over the preference of a pro-EU or pro-Russian foreign policy orientation is the least polarizing issue in Serbia. the least democratic(mean score 3.9) and they are also the least satisfied with how democracy works(mean score 3.5). It thus does not come as a surprise that among the seven countries under investigation Serbian citizens most strongly endorse if experts, not the government, make decisions for their country: 74% of them find this either“very good” or“fairly good.” “How democratic do you find the practice of opposition parties organizing protests against the government?” “How democratically is your country being governed today?” POL 25% SRB SWE 20% GER ESP 15% UKR EST 10% 5% 1 2 3 4 5 6 7 8 9 10 totally undemocratic totally democratic 20% 15% SWE GER 10% POL SRB ESP 5% UKR EST 1 2 3 4 5 6 7 8 9 10 totally undemocratic totally democratic -7.9% This is the average punishment corresponding to the share of Serbian voters willing to defect from an otherwise favored candidate once they adopt an undemocratic position. WHAT KIND OF UNDEMOCRATIC BEHAVIOR DO SERBS PUNISH? Overall, Serbian voters showed willingness to punish undemocratic behavior. The average punishment is a loss of-7.9% of the overall vote share for the candidate who adopts an undemocratic position, which is slightly above the average punishing rate across the entire sample. The undemocratic positions punished most strongly are“prosecuting journalists” with a -12.5% vote share loss and“violently disrupt opponent’s rallies” with a-11.7% vote share loss. On the other hand, the Serbs punish the items“ban of foreign union funding” and“partisan infrastructure spending” the least, as the candidates advocating for these positions lose only-4.8% and-5% of the vote share respectively. PARTY LOYALTY AND POLICY INTERESTS: WHICH ISSUES ARE MORE IMPORTANT THAN UPHOLDING DEMOCRATIC STANDARDS? Serbian voters forgive undemocratic behavior of a candidate when they come from their favorite party. They do the same when it comes to identity-based interests, but rather not for socio-economic interests. This is particularly the case with partisan interests: A candidate who comes from the respondent’s favorite party and holds an undemocratic position is nevertheless compensated. Accordingly, they are not be punished, but rewarded, and end up with a+15.6% vote share increase. This percentage is followed by identity-based interests: An undemocratic candidate who at the same time represents a voter’s favorite policy on same-sex couples’ rights or foreign policy orientation(pro-European or pro-Russian) is still rewarded+5.3% and +4.8%(final vote share) respectively. The same applies to representing a voter’s favorite language policy, with albeit a significantly lower final rewarding rate, namely+1.3%. This is also the lowest rate among those four countries in our sample in which the language issue plays a 59

How democratically elected politicians get away with autocratizing their country salient role in political processes. Contrary to our findings on identity-based interests, Serbs appear much less willing to forgive undemocratic behavior for their favorite policies on the environment and education. The undemocratic candidate ends up being rewarded with only+0.1% vote share increase in the case of the former, while they are punished with a final loss of-1.5% in the case of the latter. A greater salience of environmental issues could perhaps be explained by the fact that Serbia recently witnessed country-wide demonstrations against the exploitation of lithium that were organized by environmental organizations. ARE SUPPORTERS OF A CERTAIN POLICY INTEREST 28 MORE FORGIVING OF UNDEMOCRATIC BEHAVIOR THAN OTHERS? If we divide respondents along a left-leaning/ right-leaning axis, we can identify more nuanced findings about the willingness to punish undemocratic behavior. Concretely, when it comes to same-sex couples’ rights, right-leaning voters show less attachment to democratic principles. Contrary to left-leaning voters, who in the end still punish with a-3.2% vote share loss, right-leaning voters forgive undemocratic behavior for their favorite policy on this issue with a final+5.6% vote share increase. We can observe a similar correlation when it comes to the language issue as well: voters who prefer the use of only the Serbian language when communicating with state offices in the end punish an undemocratic candidate, albeit with a minimal -0.2% vote share loss; voters who support the right to use the Albanian language do forgive undemocratic behavior, with the candidate ending up with a+4.2% vote share increase. When it comes to foreign policy, all respondents show a willingness to forgive undemocratic behavior. This is particularly the case with voters who want a closer relationship with Russia, as this scenario results in a+12% final vote share increase – that is+10,8% more when compared with voters who demand a stronger attachment to the EU. Moreover, with respect to environment the electorate is divided, as only those who prefer lowering taxes on coal and gas even if it results in lesser use of solar and wind energy forgive undemocratic behavior for it(+0.1% final vote share increase). 6.8 Serbs see losing an election as an inherent part of democracy: The statement‘The prime minister conceded a narrow election’ got the second highest mean score of all countries under investigation. Existence of trade-offs/compensation for undermining democracy punishment reward Average punishment -7.9% Partisan Same-sex couples’ rights EU-Russia Language Environment Education Defense policy Immigration -8% +23.5% +13.2% 5.3% +12.7% 4.8% +9.2% 1.3% +8.0% 0.1% +6.4% -1.5% N/A N/A -4% 0%+4%+8%+12% Final reward 15.6% 28 Please find an overview of all tested policy positions on pg. 13. 60

IDENTITY, PARTISANSHIP, POLARIZATION WHAT ROLE DOES POLARIZATION PLAY? When it comes to same-sex couples’ rights, we can observe a correlation between an increased polarization and the readiness to forgive undemocratic behavior for this interest. Indeed, the matter of rights for same-sex couples is the most polarizing issue in Serbia. Moreover, we can observe the same correlation regarding environmental and language issues as well, albeit with a lesser intensity. With respect to the foreign policy orientation of a candidate, however, the mentioned correlation cannot be observed. As a matter of fact, the EU or Russia foreign policy orientation is the least polarizing policy issue in Serbia, as a great majority of Serbian citizens agree that the country should have close relations with both actors. This fact greatly explains the reluctance of the Serbian leadership to join the Western sanctions on Russia amid its aggression on Ukraine. Foreign policy: shift in punishment when presented with favored policy punishment reward -8.5% +9.7% 1.2% Pro-EU -9.6% +14.7% 5.1% Neutral -1.5% +13.5% 12.0% Pro-Russian -8%-4% 0%+4%+8%+12% WHO ARE THE MAIN‘DEFENDERS OF DEMOCRACY’ IN SERBIA? Average electoral punishment by respondents’ age groups 18-29 years 30-39 40-49 50-65 66-75 -2% -4% Serbia* -6% -8% -10% -12% All countries* average -14% *No data was collected for age group 66-75 in Serbia and Ukraine. As in all other countries under investigation, women punish undemocratic behavior more than men. As in many other countries, other main defenders of democracy are people who show more interest in politics and who are democratically competent, as well as those who lack an authoritarian personality. Also people who work full- and parttime punish undemocratic behavior more than comparative sub-group members. A more surprising result concerns an interesting inversed generational difference that we find in Serbia, but not in most of the other countries: pensioners and the elderly punish un democratic behavior more than the young and students. This holds true for both of the younger cohorts in our sample(18-29 and 30-39); the significant shift starts in the 40-49 cohort. In line with the average for all cases, the lower the level of education the less Serbs punish undemocratic behavior. Other groups that are less willing to punish politicians who violate democratic principles are people who are taking care of family members or were temporarily laid off, but also those with a more positive economic assessment of the country and of their family. 61

How democratically elected politicians get away with autocratizing their country WHICH PARTY‘S VOTERS PUNISH UNDEMOCRATIC BEHAVIOR THE MOST? When it comes to party affiliation, voters of all key Serbian parties show a general willingness to punish undemocratic politicians. However, there are strong differences among them: The voters of the party that dominates the Serbian party system – the Serbian Progressive Party(SNS) – are least willing to punish undemocratic behavior. The average punishment rate among SNS voters is only-4.1%. However, they do not show a willingness to punish all undemocratic positions, as“disciplining judges who criticize the government” is indeed rewarded by a+14.8% vote share increase. Looking again at punishment rates, the SNS voters are followed by the voters of the far-right Dveri with a punishment rate of-7.4%. Voters of the Serbian Socialist Party(SPS), the Democratic Party(DS) and Dosta je Bilo (It is enough) showed a very similar willingness to punish violations of democratic principles. The punishing rate of the voters of the SPS and the DS is-8.1% and-8.2% respectively, while for Dosta je Bilo it is-8.9%. Voters of the Party of Freedom and Justice(SSP) punish the most, with a punishing rate of-14.3%. In other words, the voters of the opposition parties show more attachment to democracy than the voters of the ruling SNS and SPS. Average punishment for undermining democracy punishment Demokratska SSP Dosta je bilo Stranka SPS -8.9% -8.2% -8.1% -14.3% Dveri -7.4% SNS -4.1% -2% -4% -6% -8% -10% -12% -14% Summary for Serbia The majority of Serbs value the various facets of democracy. However, they are also quite critical of the democratic system, and especially of the way democracy functions in their own country. Compared to other countries, they regard electoral fairness as particularly important. When it comes to defending democratic standards in our experiment, Serbs score even slightly higher than the average. They are most likely to defend the right of assembly and electoral fairness as well as the freedom of the press. As in all other countries under investigation, party loyalty is the most important determinant of voting behavior. Serbs completely forgive a candidate their undemocratic behavior if they are from their favorite party. Group loyalty and identity issues play the second most important role when it comes to voting decisions in Serbia. These include issues such as foreign policy orientation, language, as well as an extension of rights for same-sex couples, which are prioritized at the ballot box over rejecting undemocratic positions. For the most part, the willingness to forgive undemocratic behavior for particular interests correlates with the degree of polarization for these issues. One important exception is a preferred foreign policy position: foreign policy is not a polarizing issue in Serbia, but Serbs regard it as important. Hence, the salience of the issue could perhaps serve as one of the explanations for the tolerance of undemocratic behavior. 62

IDENTITY, PARTISANSHIP, POLARIZATION Facts& Figures Population Political system Which issues let politicians get away with undemocratic behavior? Issues displayed according 8 to their importance to voters and degree of polarization ! Issues for which voters 7 forgive undemocratic behavior Education Environment ! Same-sex couples’ rights Defense policy ! Immigration ! Language 6 Varieties of Democracy regime type 5 more important (1 10) Varieties of Democracy Liberal Democracy Index ranking(2021) 4 3 0.2 0.3 0.4 0.5 0.6 more polarizing (0.0- 1.0) 79% Democratic competence of Spanish respondents Percentage of respondents who on average rate the four democratic items as more democratic than the ten undemocratic items in our survey How much are key elements of liberal democracy valued? The higher the vote loss for an undemocratic policy proposed by a political candidate, the more a certain democratic element is valued. -11.9% -6.1% -2.5% -5.5% -2.9% -7.5% -5.7% -5.7% -3.0% Ban foreign union funding Pass laws without parliamentary debate Ban foreign NGO funding Monitor social media Partisan infrastructure spending Discipline judges Ban prayers for Muslims Prosecute journalists Violently disrupt rallies -20%-15%-10%-5% 0% Average punishment across all 7 countries Min/Max punishment across all 7 countries Punishment in Spain 63

How democratically elected politicians get away with autocratizing their country Spanish parties presented in the survey English Spanish name Socialist United We Can People´s Party Vox Workers Party Spanish Partido Socialista Unidas Podemos Partido Popular name Obrero Español (UP)(PP) Vox (PSOE) European S&D GUE/NGL EPP ECR affiliation Currently governing Citizens Ciudadanos (CS) ALDE Republican Left of Catalonia– Catalonia Yes Esquerra Republicana de Catalunya– Catalunya Sí (ERC–CatSí) Greens/EFA KNOWLEDGE ABOUT AND SUPPORT FOR DEMOCRACY Spanish respondents have a good understanding of what democracy is and what it is not. They are strongly pro-democracy oriented as well, however, slightly less so than the respondents in the two other Western European countries in our survey, Germany and Sweden. Spanish people also notably expect more from democracy than it is currently delivering. Democratic competence 29 of Spanish citizens is at 79%. For the Spaniards, it is important to live in a country that is governed democratically(mean score 8.4) and to have a democratic political system(90% of them find it either“very good” or “fairly good”). Moreover, 89% of them either“strongly agree” or“agree” that democracy is better than any other form of government. In line with these outcomes, the features of an authoritarian system – such as“having a strong leader who does not have to bother with parliament or elections” and“having the army rule” – are unpopular among Spanish respondents. More than three-quarters(76.5%) of them rank the former situation either“very bad” or“fairly bad”, while 83.5% hold the same opinion about the latter. With its past experience with military dictatorship being more recent than in Germany, Spanish people might have more awareness of the danger that it poses to democracy. Thus, the statement“The military removed a corrupt president” receives on average the lowest rating just behind Sweden, 30 but still 24.8% of Spaniards find the statement rather or completely democratic. 31 Spain also stands out for giving the lowest democratic rating of all countries to the possibility that“the high court rules a government’s policy as unconstitutional.” Yet, the Spaniards find it comparatively undemocratic(significantly more so than Germans for instance) 32 if the“country’s judiciary were staffed with individuals loyal to “How strongly do you agree that democracy is better than any other form of government?” All figures in% Sw e d en 49.5 42.5 6.0 G er ma ny 47.1 44.6 6.8 Spa in 44.4 44.7 8.8 P o la n d 32.7 56.5 9.2 S er b i a 24.4 60.9 11.9 E stoni a 23.5 60.8 13.7 Uk r a ine 15.1 37.9 38.6 8.4 20% 40% 60% 80% strongly agree agree disagree strongly disagree Due to rounding, totals may not correspond with the sum of the separate figures. the governing party.” When it comes to their democracy’s performance, however, the expectations of the Spaniards seem not to be fully fulfilled: With comparatively low mean scores for both the assessment of“how democratically the country is being governed”(5.6) and the respondent’s satisfaction with“how democracy works” (4.9), the Spaniards suggest that there is much room for improvement. The dissatisfaction also manifests itself in the fact that two thirds of all Spanish respondents(the highest number after Serbia among all seven countries under investigation) finds it“fairly good” or“very good” if experts, not the government, make decisions for the country. 29 If we define a“democratically competent” respondent as someone who, on average, rates the four democratic items as more democratic than the ten undemocratic items present in the survey. 30 On a scale from 1(not at all democratic) to 10(completely democratic), the mean score is only 4.4. 31 Rating it with an 8, 9 or 10 on a scale from 1 to 10. 32 Spain’s mean score on a 1 – 10 scale is 2.6 as opposed to Germany’s 3.1. 64

IDENTITY, PARTISANSHIP, POLARIZATION Language The highest degree of polarization in Spain concerns the right to use the Catalan language when communicating with state offices. “How satisfied are you with the way democracy works in your country?” “How democratic do you find the practice of the country’s high court ruling that a government policy was unconstitutional?” 20% 15% SWE 10% POL GER SRB 5% EST UKR ESP 1 2 3 4 5 6 7 8 9 10 not satisfied at all very satisfied 25% GER POL 20% SRB SWE 15% EST UKR 10% ESP 5% 1 2 3 4 5 6 7 8 9 10 totally undemocratic totally democratic WHAT KIND OF UNDEMOCRATIC BEHAVIOR DO SPANIARDS PUNISH? Spanish voters showed the least willingness among the seven countries under investigation to punish undemocratic behavior. The average punishment is a loss of-5.9% of the overall vote share for the candidate who adopts an undemocratic position. The most punishing undemocratic positions are“violently disrupt opponent’s rallies” with an-11% vote share loss and‘partisan infrastructure spending’ with a-7.4% vote share loss. On the other hand, Spaniards punish“passing laws without parliamentary debate” and “prosecuting journalists” the least, as the candidates advocating these undemocratic positions lose only-2.4% and-2.9% of the vote share respectively. Hence, in this aspect Spain differs from the remaining countries, as it is the only country in which attacking journalists is one of the least punishable undemocratic behaviors. -5.9% punishment rate Spanish voters show the least willingness to punish undemocratic behavior of all voters in the countries under investigation. PARTY LOYALTY AND POLICY INTERESTS: WHICH ISSUES ARE MORE IMPORTANT THAN UPHOLDING DEMOCRATIC STANDARDS? Spanish voters forgive undemocratic behavior of a candidate when they are from their favorite party. They do the same when it comes to identity-based interests, but not for socio-economic interests. The results are particularly significant when it comes to partisan interests: The candidate who advocates undemocratic policies, but is from the respondent’s favorite party, is not at all punished, but instead ends up being rewarded with a+21% vote share increase. This is the highest final rewarding rate among all seven countries under investigation. Favorite identity-based policies can compensate for undemocratic behavior as well: An undemocratic candidate who also represents a voter’s favorite policy on same-sex couples’ rights, language or immigration can in the end count on a final reward of+8.4%,+5.4% and+3.1% respectively. Contrary to partisan and identity-based interests, advocating for favorite socio-economic and defense policies does not bring a final electoral reward. In other words, an undemocratic candidate ends up being punished despite advocating for not only a favorite policy regarding education(-0.2% vote share loss) and the environment(-0.5% vote share loss), but a favorite defense policy as well(-1.4% vote share loss). 65

How democratically elected politicians get away with autocratizing their country ARE SUPPORTERS OF A CERTAIN POLICY INTEREST 33 MORE FORGIVING OF UNDEMOCRATIC BEHAVIOR THAN OTHERS? If we divide respondents along a left-leaning/right-leaning axis, we can identify more nuanced findings about the willingness to punish undemocratic behavior. Concretely, when it comes to same-sex couples’ rights, voters with a more progressive view show less attachment to democratic principles. Indeed, only those who would deny same-sex couples the right to marry and adopt children would not forgive a candidate their undemocratic behavior. They would in the end still punish that candidate with a-2.2% vote share loss, whereas left-leaning voters would forgive undemocratic behavior for their favorite policy on this issue, with a remarkable final+10% vote share increase. When it comes to immigration, we can see that centrist voters value democracy more than voters on the fringes. Although centerleft and center-right voters still forgive undemocratic behavior for their favorite immigration policy, the final rewarding rates – +0.2% and+0.4% vote share increase respectively – are not significant. Contrary to this, right-wing voters(+3.9% final vote share increase) and particularly left-wing voters(+6% final vote share increase) show a clear willingness to disregard democratic principles for their immigration-related interests. A stronger willingness of the left-leaning voters to forgive undemocratic conduct can also be observed with respect to a candidate’s position on the language issue. Voters of candidates who advocate for the use of the Catalan language when communicating with state offices forgive undemocratic behavior for it. Such an undemocratic candidate ends up with a+13.2% vote share increase, which is 11.8% more than in the case of a candidate who advocates for the use of the Spanish language only. Existence of trade-offs/compensation for undermining democracy punishment reward Average punishment -5.9% Partisan Same-sex couples’ rights Language Immigration Education Environment Defense policy EU-Russia +26.9% +14.3% 8.4% +11.3% 5.4% +9.0% 3.1% +5.7% -0.2% +5.4% -0.5% +4.5% -1.4% N/A -4% 0%+4%+8%+12%+16% Final reward 21.0% WHAT ROLE DOES POLARIZATION PLAY? In Spain, the highest degree of polarization concerns the issue of language. As we have seen above, this is also an issue for which voters disregard undemocratic behavior. Hence, we can observe a correlation between an increased polarization and the willingness to forgive undemocratic behavior. The same holds true for the issue of immigration. 33 Please find an overview of all tested policy positions on pg. 13. Contrary to these findings, the issue of same-sex couples’ rights appears not to be a polarizing one, yet the Spaniards still tolerate undemocratic behavior for it. 66

IDENTITY, PARTISANSHIP, POLARIZATION 2/ 3 Two-thirds of all Spaniards find it ‘fairly good’ or‘very good’ if ‘experts, not the government, make decisions for the country.’ Language: shift in punishment when presented with favored policy punishment reward -5.3% +6.7% 1.4% Pro-Spanish language -7.1% -4% +20.3% 13.2% Pro-Catalan language 0%+4%+8%+12% Average electoral punishment by respondents’ age groups 18-29 years 30-39 40-49 50-65 66-75 -2% -4% -6% -8% -10% -12% Spain All countries* average -14% *No data was collected for age group 66-75 in Serbia and Ukraine. WHO ARE THE MAIN‘DEFENDERS OF DEMOCRACY’ IN SPAIN? As in most of the other countries under investigation, women punish undemocratic behavior more than men. The same holds true for those who showed more interest in politics, those who are democratically compe tent, those who are less religious, those who are more trusting of other people and those who lack an authoritarian personality. The youngest voter group(18-29 years of age) in Spain punishes undemocratic behavior slightly more than the country sample’s average of the same age group. Spain’s 50 to 65-year-olds punish about as much as the young and are almost on par with the average. All other age groups punish politicians who violate democratic principles distinctly less than the sample’s average. When it comes to education, the more educated punish slightly more, but just like in Germany and Sweden, the level of education does not have a great effect on the degree of punishing undemocratic behavior, especially not when comparing the results for voters with a medium and high level of education. Looking at the occupation of voters, Spain stands out as all of the following groups punish undemocratic behavior equally: people working full-time or part-time, the retired, the unemployed and people taking care of their family. Only students punish distinctly more. WHICH PARTY‘S VOTERS PUNISH UNDEMOCRATIC BEHAVIOR THE MOST? When it comes to party affiliation, voters of all Spanish parties showed a general willingness to punish undem ocratic politicians. While one can observe variations, they can hardly be described as significant: The voters of the conservative PP and the far-right Vox punish undemocratic behavior the least, as the average punishment rate among them is only-4.9%(vote share loss). An almost identical punishment rate also characterizes the 67

How democratically elected politicians get away with autocratizing their country voters of the ruling social-democratic PSOE(-5%) and voters of Catalan ERC(-5.5%). Although they punish on average, the PSOE voters actually reward violations of some democratic principles such as prosecuting journalists with+10.7% and disciplining judges with a+11.2% vote share increase. Voters of the far-left Unidas Podemos and the liberal Ciudadanos show more willingness to punish undemocratic behavior as the vote share loss is-7.2% and-7.5% respectively. However, the voters of Podemos also reward one violation of democratic principles – passing laws without parliamentary debate – with a+9.3% vote share increase. Interestingly, respondents who do not vote for either of these six parties punish the most, with an-8.6% punishment rate. Average punishment for undermining democracy Ciudadanos Unidas Podemos ERC PSOE Vox PP punishment -7.5% -7.2% -5.5% -5.0% -4.9% -2% -4% -4.9% -6% Summary for Spain A large majority of Spanish respondents value the various facets of democracy. However, they are quite critical of the democratic system, and especially of the way democracy functions in their own country. Compared with other countries, they regard protecting democracy from a strong leader and governing influence of the military as highly important. Even though Spaniards express their high regard for democratic standards and the rule of law in their responses to the direct questions, in our experiment, they show a lower willingness than respondents in all other countries to actually punish undemocratic behavior at the ballot box. They are most likely to defend electoral fairness. As in all other countries under investigation, party loyalty is the most important determinant of voting behavior. Spaniards completely forgive a candidate their undemocratic behavior if they are from their favorite party. Moreover, party loyalty is strongest in Spain when contrasting it to all other countries under investigation. Group loyalty and identity issues play the second most important role when it comes to voting decisions in Spain. These include issues such as rights for same-sex couples, language, as well as immigration, which are prioritized at the ballot box over rejecting undemocratic positions. Polarization plays a role as well, as the willingness to forgive undemocratic behavior for particular interests mostly, but not always, correlates with the degree of polarization for these issues. Albeit not being a polarizing issue, the fact that the Spaniards rate an extension of rights for same-sex couples as very important could be one of the reasons for a high tolerance of undemocratic behavior in relation to this issue. 68

IDENTITY, PARTISANSHIP, POLARIZATION Facts& Figures Population Political system Varieties of Democracy regime type Which issues let politicians get away with undemocratic behavior? Issues displayed according 8 to their importance to voters and degree of polarization ! Issues for which voters 7 forgive undemocratic behavior Education ! ! Same-sex couples’ rights Environment ! Immigration 6 Defense policy 5 more important (1 10) Varieties of Democracy Liberal Democracy Index ranking(2021) 4 3 0.2 0.3 0.4 0.5 0.6 more polarizing (0.0- 1.0) 85% Democratic competence of Swedish respondents Percentage of respondents who on average rate the four democratic items as more democratic than the ten undemocratic items in our survey How much are key elements of liberal democracy valued? The higher the vote loss for an undemocratic policy proposed by a political candidate, the more a certain democratic element is valued. -6.3% -4.9% -5.1% -8.1% -10.4% -13.7% -14.5% -10.0% -17.3% Ban foreign union funding Pass laws without parliamentary debate Ban foreign NGO funding Monitor social media Partisan infrastructure spending Discipline judges Ban prayers for Muslims Prosecute journalists Violently disrupt rallies -20%-15%-10%-5% 0% Average punishment across all 7 countries Min/Max punishment across all 7 countries Punishment in Sweden 69

How democratically elected politicians get away with autocratizing their country Swedish parties presented in the survey English name Swedish name Swedish Social Democratic Party Sveriges Socialdemokratiska arbetarparti(S/SAP) Green Party Miljöpartiet de Gröna(MP) European S&D affiliation Greens/EFA Currently governing Left Party Vänsterpartiet (V) Moderate Party Sweden Democrats Centre Party Moderaterna (M) Sverigedemokraterna(SD) Centerpartiet(C) GUE/NGL EPP ECR ALDE KNOWLEDGE ABOUT AND SUPPORT FOR DEMOCRACY Among all seven European countries in our sample, Sweden seems to be the democratic champion. Some findings however deserve further observation in the future to ensure that Sweden remains a well-functioning democracy. Across the entire sample, the respondents from Sweden have the best understanding of what democracy is(and what it is not) and they also showed the greatest support for democracy and its principles. Indeed, democratic competence of Swedish citizens is 85%. 34 While 20.9% of Swedes find it rather or completely democratic 35 if“The military removed a corrupt president”, this is still the lowest share of people agreeing to that statement compared to the other six countries. For the Swedes, it is furthermore important to live in a country that is governed democratically(mean score 8.8) and to have a democratic political system(93% find it either“very good” or“fairly good”). They also clearly find democracy better than any other form of government as 92% of them either“strongly agree” or“agree” with that statement. Moreover, the features of an authoritarian system – such as“having a strong leader who does not have to bother with parliament” or“having the army rule” – are highly unpopular with the Swedish population. Indeed, 85% of Swedes find the former situation and 88% find the latter situation either“very bad” or“fairly bad.” “How democratic do you find the practice of the military removing a corrupt President?” 20% SRB 15% POL ESP 10% UKR GER EST 5% SWE 1 2 3 4 5 6 7 8 9 10 totally undemocratic totally democratic “How democratic do you find the practice of the country’s high court ruling that a government policy was unconstitutional?” 25% GER POL 20% SRB SWE 15% EST UKR 10% ESP 5% 1 2 3 4 5 6 7 8 9 10 totally undemocratic totally democratic Democracy Swedes show the greatest support for democracy and its principles. 34 If we define a“democratically competent” respondent as someone who, on average, rates the four democratic items as more democratic than the ten undemocratic items present in the survey. 35 Rating it with an 8, 9 or 10 on a scale from 1 to 10. 70

IDENTITY, PARTISANSHIP, POLARIZATION -10.0% Swedish voters show the greatest willingness to punish undemocratic behavior, with a-10% punishment rate. There were only two questions where Swedes were not among the most democratic respond ents: Most notably, they ranked the statement “The prime minister conceded a narrow elec tion” least democratic when comparing the Swedish score to all other countries(with a mean score of 6.1 on a scale from 1 to 10). They also rated the statement“The high court rated a government policy unconstitutional” relatively low(with a mean score of 5.9), on par with Poland and Estonia and behind Serbia and Germany. The Swedes appear to be more satisfied with how their democracy performs than the respondents of other countries. Across the entire sample, the Swedes see their country as being governed the most democratic(mean score 6.7.) and they are also the most satisfied with how their democracy works(mean score 6.1). However, the level of satisfaction can hardly be described as high. This is also manifested in the fact that 44% of Swedish respondents(albeit the lowest number among seven countries) find it“fairly good” or“very good” if experts, not the government, make decisions for the country. ARE SWEDISH VOTERS WILLING TO PUNISH UNDEMOCRATIC BEHAVIOR? Swedish voters showed the greatest willingness among the seven countries under investigation to punish undemocratic behavior. The average punishment is a loss of-10.0% of the overall vote share for the candidate who adopts an undemocratic position. The most punishing undemocratic positions are“violently disrupt opponent’s rallies” which results in a-17.2% vote share loss and“prosecuting journalists” with a -14.4% vote share loss. On the other hand, the Swedes punish“passing laws without parliamentary debate” and“ban foreign NGO funding” the least, as the candidates advocating those undemocratic positions lose only-4.8% and-5.0% of the vote share respectively. PARTY LOYALTY AND POLICY INTERESTS: WHICH ISSUES ARE MORE IMPORTANT THAN UPHOLDING DEMOCRATIC STANDARDS? Swedish voters forgive undemocratic behavior of a candidate when they are from their favorite party. They do the same when it comes to identity-based interests, and partly for socio-economic interests, but not for defense policy interests. This is particularly the case with Existence of trade-offs/compensation for undermining democracy punishment reward Average punishment -10.0% Partisan Same-sex couples’ rights Immigration Environment Education Defense policy EU-Russia Language +21.5% +15.8% 5.8% +14.2% 4.2% +11.8% 1.8% +8.2% -1.8% +8.2% -1.8% 7.4% N/A N/A -8%-4% 0%+4%+8% Final reward 11.5% 71

How democratically elected politicians get away with autocratizing their country partisan interests, as the undemocratic candidate from the respondent’s favorite party is compensated for that mere fact. Hence, they are not punished, but end up being rewarded with a+11.5% vote share increase, which is, however, the lowest rewarding rate among the seven countries under investigation. Partisan interests are followed by identity-based interests. The undemocratic candidate with a favorite policy on same-sex couples’ rights or immigration can count on a final+5.8% and+4.2% vote share increase respectively. In the case of immigration, this is the highest rewarding rate among the five countries under observation in which this issue is salient. That means that there is a higher chance than in Estonia, Spain, Germany and Poland that the position on immigration of a candidate has the potential to be instrumentalized for an unpunished undemocratic behavior. The Swedes also show willingness to forgive undemocratic behavior for a favorite policy on the environment, albeit to a lower degree: the final rewarding rate is+1.8%. Notably, the willingness to punish disappears when it comes to a voter’s favorite policy regarding education. Despite advocating for it, an undemocratic candidate is still punished by losing-1.8% of the vote share in the end. In addition, there is a lack of willingness to reward undemocratic candidates who advocates for a voter’s favorite defense policy as well. The undemocratic candidate ends up losing-1.8% of the vote share. ARE SUPPORTERS OF A CERTAIN POLICY INTEREST 36 MORE FORGIVING OF UNDEMOCRATIC BEHAVIOR THAN OTHERS? If we divide respondents along a left-leaning/right-leaning axis, we can identify more nuanced findings about the willingness to punish undemocratic behavior. Specifically, when it comes to the issue of same-sex couples’ rights, those on the fringes of the political spectrum show less attachment to democratic principles. While the final rewarding rate among the voters on the right is+4.5%, among the left-leaning voters it is even higher at+6.9%. When it comes to immigration, only center-left voters do not forgive undemocratic behavior, as they still punish with a considerable final -5.6% vote share loss. The rest of the electorate acts otherwise. The strongest tendency can thereby be observed within the group of center-right voters who reward an undemocratic candidate for their favorite immigration policy 37 with a final+12.5% vote share increase. This is the highest final rewarding rate across entire sample. When it comes to the issue of environment, right-leaning voters seem to be much more forgiving(the undemocratic candidate would end up with a+4.1% vote share increase) than left-leaning ones(+0.4% vote share increase). Polarization matters Swedes show a willingness to forgive a candidate’s undemocratic position in all issues that are polarizing in Swedish society. WHAT ROLE DOES POLARIZATION PLAY? In the case of Sweden, we can indeed observe a wide-ranging correlation between the increased polarization along some issues and the willingness to forgive undemocratic behavior for related interests. This foremostly applies to the issue of immigration, followed by samesex couples’ rights and environmental issues. 36 Please find an overview of all tested policy positions on pg. 13. 37 “Ban immigration from outside of the EU.” 72

IDENTITY, PARTISANSHIP, POLARIZATION Average electoral punishment by respondents education Level of education low medium high -2% -4% -6% -8% -10% All countries average Sweden Immigration: shift in punishment when presented with favored policy punishment reward -16.1% +19.3% -10.3% +4.7% -5.6% 3.2% Left CenterLeft -5.9% +18.4% +6.3% -2.6% 3.7% 12.5% CenterRight Right -12%-8%-4% 0%+4%+8%+12% WHO ARE THE MAIN‘DEFENDERS OF DEMOCRACY’ IN SWEDEN? Young people and women punish undemocratic behavior more than the rest of the respective sub-group. Interestingly, part-time workers punish more than full-time employees (but perhaps that is due to a higher share of women working part-time). In terms of attitudes, the results are not surprising: Those who show an increased interest in politics and have higher levels of political knowledge, who are democratically competent and more trustful, as well as less religious people and those who lack an authoritarian personality show a greater willingness to punish undem ocratic behavior. When it comes to education, the more educated punish slightly more. But just like in Spain and Germany, the level of education does not have a great effect on the degree of punishing undemocratic behavior, especially not between medium and high levels of education. Who then shows less willingness to punish politicians who violate democratic principles? The unemployed together with those taking care of family members and those with a more negative economic assessment of their country and their family. WHICH PARTY‘S VOTERS PUNISH UNDEMOCRATIC BEHAVIOR THE MOST? When it comes to party affiliation, voters of all Swedish parties show a general willingness to punish undem ocratic politicians. Yet one can also observe great variations: The voters of the far-right Swedish Democrats (SD) punish undemocratic behavior the least, as the punishment rate 73

How democratically elected politicians get away with autocratizing their country among them is only-4.5% of vote share loss. Also, the willingness to punish does not apply to all undemocratic positions, as SD voters reward the“ban on prayers for Muslims” with a+7% vote share increase. The voters of the parties that are playing the most important role in the Swedish political system – the Social-Democrats (S) and the Moderates(M) – show more willingness to punish violations of democratic principles with a-10% and a-10.8% punishing rate respectively. However, voters of the SocialDemocrats reward“passing laws without parliamentary debate” with+7.5%. The voters of the far-left party Vänsterpartiet and the voters of the Centre Party punish identically –-13.5% –, while the voters of the Greens punish the most with a -20.9% punishing rate. Interestingly, the punishing rate of those who do not vote in the elections is by far the lowest:-2.6%. Average punishment for undermining democracy punishment Miljöpartiet de Social- SverigeGröna Centerpartiet Vänsterpartiet Moderaterna demokraterna demokraterna -13.5%-13.5% -10.8% -10.0% -20.9% -4.5% -2% -4% -6% -8% -10% -12% -14% -16% -18% -20% Center-right voters& immigration Center-right voters show the greatest willingness to reward an undemocratic candidate who also proposes the voters’ favorite immigration policy with a final +12.5% vote share increase. Summary for Sweden A large majority of respondents in Sweden support democratic standards and the rule of law. Across the entire sample, Swedes are also the most satisfied with how their democracy works. Swedes express their high regard for democracy not only in their responses to the direct questions, but also in our candidate choice experiment: the Swedes punish undemocratic behavior at the ballot box more strongly than respondents of all other countries. They are most likely to defend the right of assembly and electoral fairness as well as the freedom of the press. As in all other countries under investigation, however, party loyalty is the most important determinant of voting behavior. Swedes completely forgive a candidate their undemocratic behavior if they are from their favorite party. In contrast to other countries under investigation, we can see that all issues for which Swedes show a willingness to forgive undemocratic behavior are the ones that are polarizing in Swedish society. These include identity issues(rights for same-sex couples and immigration) as well as one socio-economic issue: environmental policy. Hence, identity issues and, even more so, polarization play an important role when it comes to punishing undemocratic behavior in Sweden. 74

IDENTITY, PARTISANSHIP, POLARIZATION Facts& Figures Population Political system Which issues let politicians get away with undemocratic behavior? Issues displayed according 8 to their importance to voters and degree of polarization ! Issues for which voters 7 forgive undemocratic behavior Environment ! Education ! EU-Russia ! Language 6 Varieties of Democracy regime type 5 more important (1 10) Varieties of Democracy Liberal Democracy Index ranking(2021) 4 3 0.2 Same-sex ! couples’ rights 0.3 0.4 0.5 0.6 more polarizing (0.0- 1.0) 77% Democratic competence of Ukrainian respondents Percentage of respondents who on average rate the four democratic items as more democratic than the ten undemocratic items in our survey How much are key elements of liberal democracy valued? The higher the vote loss for an undemocratic policy proposed by a political candidate, the more a certain democratic element is valued. -3.8% -4.0% -4.2% -6.6% -8.3% -2.8% -4.2% -5.8% -10.6% Ban foreign union funding Pass laws without parliamentary debate Ban foreign NGO funding Monitor social media Partisan infrastructure spending Discipline judges Ban prayers for Muslims Prosecute journalists Violently disrupt rallies -20%-15%-10%-5% 0% Average punishment across all 7 countries Min/Max punishment across all 7 countries Punishment in Ukraine 75

How democratically elected politicians get away with autocratizing their country Ukrainian parties presented in the survey English name Ukrainian name European affiliation Currently governing Servant of the People Слуга народу (SN) ALDE European Solidarity Європейська солідарність (YeS) EPP(Observer) All-Ukrainian Union "Fatherland" ВО « Батьківщина» (VOB) EPP(Observer) Opposition Platform for Life Опозицiйна платформа – За життя Voice Голос (PH) ALDE KNOWLEDGE ABOUT AND SUPPORT FOR DEMOCRACY The Ukrainian people have a good understanding of what democracy is and what it is not. This is confirmed by their democratic competence that is at 77%. Moreover, a majority of Ukrainian citizens are pro-democracy oriented, however, some items related to the support for democracy enjoy significantly less support than in the other six European countries under investigation. Indeed, for Ukrainians it is important to live in a country that is governed democratically(mean score 8.4) and to have a democratic political system(80% of Ukrainians rank this as either“very good” or“fairly good”). Yet, only a slim majority of Ukrainians(53%) either“strongly agree” or“agree” that democracy is better than any other form of government. The population seems to be really torn on this question. When we look at the Ukrainians’ attitude toward some features of an authoritarian system, we find a strong rejection of the item“having the army rule” as more than three-quarters (76.4%) of Ukrainians find this proposition either“very bad” or “fairly bad.” If“the military removed a corrupt president”, 23.3% of Ukrainians would find this rather or completely democratic 39 - a similar outcome as in other countries. Still, on average, Ukrainians rank just third behind Swedes and Spaniards in rating this act as undemocratic. 40 Moreover, the country is divided over the acceptance of having a strong leader who does not have to bother with parliament or elections: just 50% find this situation either“very bad” or“fairly bad.” The uncertainties in the last few years especially after the annexation of Crimea – a development that tends to favor the population’s acceptance of a “How strongly do you agree that democracy is better than any other form of government?” All figures in% Sw e d en 49.5 42.5 6.0 G er ma ny 47.1 44.6 6.8 Spa in 44.4 44.7 8.8 P o la n d 32.7 56.5 9.2 S er b i a 24.4 60.9 11.9 E stoni a 23.5 60.8 13.7 Uk r a ine 15.1 37.9 38.6 8.4 20% 40% 60% 80% strongly agree agree disagree strongly disagree Due to rounding, totals may not correspond with the sum of the separate figures. concentration of power – have most likely contributed to some of these outcomes. In comparison, Ukrainians seem to highly appreciate some civil liberties: Journalists who frequently disagree with the president’s policies, for instance, are comparatively more appreciated than in other countries, 41 more even than in Sweden. If journalists were prosecuted for criticizing the government, however, Ukrainians turn out to be the least critical of this practice. When it comes to the performance of their democracy, Ukrainians seem to demand and expect more. They appear to be less satisfied with the state of democracy in their country, 38 Please note that our survey was conducted at the end of 2021, hence prior to the 2022 start of the unprovoked Russian war against Ukraine. We are aware that some positions or views might have shifted in light of the past months. However, we decided to nevertheless report our findings as they might still be providing a valuable baseline for understanding democracy and the rule of law in Ukraine, particularly with regard to a possible EU-membership. 39 Rating it with an 8, 9 or 10 on a scale from 1 to 10. 40 On a scale from 1(not at all democratic) to 10(completely democratic), the mean score is 4.4. 41 Reaching a mean rating of 6.7 on a scale from 1 to 10. 76

IDENTITY, PARTISANSHIP, POLARIZATION “How good or bad do you find having a strong leader who does not have to bother with parliament or elections?” All figures in% Sw e d en G er ma n y Spa in Pola n d S er b i a E s to ni a 64.9 20.1 11.6 3.4 56.8 25.1 14.9 3.2 55.3 21.3 17.1 6.3 42.7 32.2 19.9 5.1 39.6 32.5 20.8 7.1 26.4 34.4 32.3 7.0 Uk r a ine 20.1 30.0 34.9 15.0 v er y bad 20% 40% 60% fa ir ly bad fa ir ly g ood 80% v er y g ood Due to rounding, totals may not correspond with the sum of the separate figures. “How democratic do you find the practice of journalists frequently disagreeing with the president's policies?” POL 25% SWE SRB 20% GER UKR 15% ESP EST 10% 5% 1 2 3 4 5 6 7 8 9 10 totally undemocratic totally democratic i.e., how democratically the country is being governed(mean score 4.8), and with how democracy works in practice(mean score 4.2) as well. Such a sentiment is also manifested in the fact that half of the population finds it“fairly good” or“very good” that experts, not the government, make decisions for the country. Judiciary Ukraine is the only country in which attacks on the judiciary is one of the least punishable undemocratic behaviors. WHAT KIND OF UNDEMOCRATIC BEHAVIOR DO UKRAINIANS PUNISH? Ukrainian voters showed willingness to punish undemocratic behavior. Yet the average punishment – a loss of-7.1% of the overall vote share for the candidate who adopts an undemocratic position – is the second lowest across the entire sample. The most punishing undemocratic positions are“violently disrupt opponent’s rallies”with a-10.5% vote share loss and“partisan infrastructure spending” with a-8.3% vote share loss. On the other hand, Ukrainians punish “passing laws without parliamentary debate” and“disciplining judges who publicly criticize the government” the least as the candidates advocating for these undemocratic positions lose only-0.3% and-2.8% of the vote share respectively. In this aspect, Ukraine differs from the other countries under investigation, as it is the only country in which“attacks on the judiciary” is one of the least punishable undemocratic behaviors. Existence of trade-offs/compensation for undermining democracy punishment reward Average punishment -7.1% Final reward Partisan +25.1% 18.0% EU-Russia +16.4% 9.3% Language +13.9% 6.8% Same-sex couples’ rights +12.2% 5.1% Environment +7.3% 0.2% Immigration +7.1% 0.0% Education +5.1% -2.0% Defense policy N/A -8%-4% 0%+4%+8%+12%+16% 77

How democratically elected politicians get away with autocratizing their country PARTY LOYALTY AND POLICY INTERESTS: WHICH ISSUES ARE MORE IMPORTANT THAN UPHOLDING DEMOCRATIC STANDARDS? Ukrainian voters forgive undemocratic behavior of the candidate when they come from their favorite party. They do the same for identity-based interests and partly for socio-economic interests as well. As in all the other countries under investigation, this is particularly the case with partisan interests. Hence, a candidate from the respondent’s favorite party who advocates undemocratic policies is not punished, but instead ends up being rewarded with a+18% vote share increase. Favorite identity-based policies can compensate for undemocratic behavior as well. While for other countries in the sample the favorite policy on same-sex couples’ rights is accompanied by the highest final rewarding rate when it comes to identity-based interests, for Ukraine, however, this is the case with the voter’s favorite policy on the country’s foreign policy orientation(pro-European or pro-Russian) – here we find a+9.3% vote share increase. This is followed by the voter’s favorite policy on the language issue and then same-sex couples’ rights, with a+6.8% (the highest among the four countries in which this issue is salient) and a+5.1% final rewarding rate respectively. This is in accordance with the dominance of national identity debates in the Ukrainian political system. Ukrainians also show willingness to sacrifice democracy for their favorite environmental policy, albeit to a much lower degree: the final rewarding rate is only+0.2%. This, moreover, does not seem to be the case when it comes to a favorite policy regarding education. Despite advocating for it, the undemocratic candidate is still punished and loses-2% of the vote share. ARE SUPPORTERS OF A CERTAIN POLICY INTEREST 42 MORE FORGIVING OF UNDEMOCRATIC BEHAVIOR THAN OTHERS? -7.5% This is the average punishment corresponding to the share of Ukrainian voters willing to defect from an otherwise favored candidate once they adopt an undemocratic position. If we divide respondents along a left-leaning/ right-leaning axis, we can identify more nuanced findings about the willingness to punish undemocratic behavior. Concretely, the advocates of closer relations with the EU at the expense of a closer relationship with Russia as well as those who want close relationship with both actors forgive undemocratic behavior of the respective candidate, with a+9.5% and a+11.2% final vote share increase respectively. Contrary to this, advocates of a stronger attachment to Russia would neither reward, nor punish for it. When it comes to the language issue we can observe a reversed trend. Proponents of the right to use Russian when communicating with state offices would clearly forgive undemocratic behavior for it. An undemocratic candidate who advocates such a policy would end up with a+13.7% vote share increase, which is 11.9% more compared to the undemocratic candidate who demands the sole use of Ukrainian in such occasions. Regarding the issue of rights for same-sex couples, those on the right show less attachment to democratic principles. Indeed, they forgive undemocratic behavior for their favorite policy on this issue and an undemocratic candi date who advocated for it ends up increasing their vote share by+8.8%. The final rewarding rate among leftist voters is smaller, yet still considerable at+4.1%. When it comes to the issue of environment, the electorate appears to be divided, as only right-leaning voters forgive undemocratic behavior for their favorite policy (+2.3% final vote share increase). 42 Please find an overview of all tested policy positions on pg. 13. 78

IDENTITY, PARTISANSHIP, POLARIZATION Language: shift in punishment when presented with favored policy punishm. reward -8.0% +9.8% 1.8% Pro-Ukrainian language +19.1% -5.4% 13.7% Pro-Russian language -4% 0%+4%+8%+12% WHAT ROLE DOES POLARIZATION PLAY? Ukraine has the highest levels of polarization of all countries under investigation. Polarization is the highest on the issue of which rights to grant to same-sex couples, followed by the foreign policy orientation(EU vs. Russia) and the issue of language use. For all these cases, we find an overlap between a high polarization and a willingness to forgive undemocratic behavior. However, there is also one non-polarizing issue for which the Ukrainians would forgive undemocratic behavior: the environmental policy. Average electoral punishment by respondents’ age groups 18-29 years 30-39 40-49 50-65 66-75 -2% -4% -6% Ukraine* -8% -10% -12% All countries* average -14% *No data was collected for age group 66-75 in Serbia and Ukraine. Average electoral punishment by respondents agreeing/disagreeing with the statement that“democracy is better than any other form of government” strongly agree agree disagree strongly disagree -2% -4% -6% -8% -10% Ukraine All countries average WHO ARE THE MAIN‘DEFENDERS OF DEMOCRACY’ IN UKRAINE? As in the other countries under investigation, women, students and the more educated among the popula tion defend democracy most. In addition, those with more political knowledge, those who work full time, those who are more trustful, those who lack an authoritarian personality and, interestingly, those who are unemployed defend democracy as well. However, in this regard we can identify some differences between Ukraine and the remaining countries in our sample: In Ukraine, those who show an increased inter est in politics, who are less religious and who are democratically compe tent do not punish more. We also found that those who agreed or strongly agreed with the statement “democracy is better than any other form of government” punish undemocratic behavior actually clearly less than those who hold the opposite view. Who else shows less willingness to punish politicians who violate democratic principles? Pensioners, citizens who got temporarily laid off, citizens with a more positive economic assessment of the country and, surprisingly, the young as well. WHICH PARTY‘S VOTERS PUNISH UNDEMOCRATIC BEHAVIOR THE MOST? When it comes to party affiliation, voters of all Ukrainian parties showed a general willingness to punish un democratic politicians, however, voters of some parties are willing to punish more than voters of others, a tendency shared among all countries under investigation.“Fatherland” voters show the least willingness to punish undemocratic behavior, with an average punishment rate of only -5.4%(vote share loss). They are followed by voters of the ruling Servant of the People, who punish slightly more with-5.7%. These particular voters also showed a readiness to reward violation of some democratic principles: disciplining judges with +16.4%, passing laws without parliamentary debate with+18.6% and prosecuting journalists with a +19.8% vote share increase. 79

How democratically elected politicians get away with autocratizing their country The voters of other Ukrainian parties punish only slightly more. The punishment rate in the case of the Opposition Platform-For Life voters is-6%. Violations of democratic principles is punished the most by voters of the Voice and the European Solidarity, with a-6.4% and-9.9% punishment rate respectively. Interestingly, the punishment rate among the respondents who would not vote for either of these parties is the second highest:-8%. Average punishment for undermining democracy European Solidarity Voice Opposition Platform for Life Servant of the People Fatherland -2% -4% Polarized Ukraine is the most polarized country across the entire sample. punishment -6.4% -6.0% -5.7% -5.4% -6% -8% -9.9% Summary for Ukraine A majority of Ukrainian citizens are pro-democracy oriented, however, some democratic standards enjoy significantly less support than in the other six European countries under investigation. Ukrainians are quite critical of the way democracy works in their own country. When it comes to defending democratic standards in our experiment, Ukrainians score below average(but with only the second lowest punishing rate behind Spain). Ukrainians are most likely to defend electoral fairness. As in all other countries under investigation, party loyalty is the most important determinant of voting behavior. Ukrainians completely forgive a candidate their undemocratic behavior if they are from their favorite party. Group loyalty and identity issues play the second most important role when it comes to voting decisions in Ukraine. These include issues such as foreign policy orientation, language, as well as rights for same-sex couples, which are prioritized at the ballot box over rejecting undemocratic positions. Ukraine has the highest level of polarization of all countries under investigation, and this higher degree of polarization mostly correlates with a higher willingness to forgive undemocratic behavior for polarizing issues. However, one exception is the issue of environmental policy: it is not a polarizing issue, but it is regarded as important by Ukrainians. Hence, the salience of the issue can perhaps serve as one explanation for the tolerance of undemocratic behavior. 80

IDENTITY, PARTISANSHIP, POLARIZATION 81

REFERENCES

IDENTITY, PARTISANSHIP, POLARIZATION LIST OF PARTIES IN THE SURVEY ECR EKRE SDE ALDE EPP Greens/ EFA GUE/ NGL ID RE S&D AfD CDU CSU FDP SPD L PiS PO European Conservatives and Reformists Party Eesti Konservatiivne Rahvaerakond – Conservative People's Party of Estonia Sotsiaaldemokraatlik Erakond – Social Democratic Party Isamaa – Pro Partia Alliance of Liberals and Democrats for Europe European People‘s Party The Greens/ European Free Alliance Estonia Estonia Estonia Estonia Europe Europe Europe European United Left/ Nordic Green Left Identity and Democracy Renew Europe Progressive Alliance of Socialists and Democrats Alternative für Deutschland – Alternative for Germany Christlich Demokratische Union Deutschlands – Christian Democratic Union of Germany Christlich-Soziale Union in Bayern – Christian Social Union in Bavaria Freie Demokratische Partei – Free Democratic Party Sozialdemokratische Partei Deutschlands – Social Democratic Party of Germany Bündnis 90/Die Grünen – Alliance 90/The Greens Die Linke – The Left Lewica – The Left Prawo i Sprawiedliwość – Law and Justice Platforma Obywatelska – Civic Platform Konfederacja – Confederation Europe Europe Europe Europe Germany Germany Germany Germany Germany Germany Germany Poland Poland Poland Poland DS SNS SPS SSP CS ERC PP PSOE UP Vox C M MP S SD V PH SN VOB YeS Polska 2050 – Poland 2050 Demokratska Stranka – Democratic Party Srpska Napredna Stranka – Serbian Progressive Party Socijalistička Partija Srbije – Serbian Socialist Party Stranka Slobode i Pravde – Party of Freedom of Justice Dveri Dosta je Bilo – It is enough Ciudadanos – Citizens Esquerra Republicana de Catalunya–Catalunya Sí – Republican Left of Catalonia– Catalonia Yes Partido Popular – People´s Party Partido Socialista Obrero Español – Spanish Socialist Workers Party Unidas Podemos – United We Can Vox Centerpartiet – Centre Party Moderaterna – Moderates Green Party – Miljöpartiet de Gröna Sveriges Socialdemokratiska Arbetarparti – Swedish Social Democratic Party Sverigedemokraterna – Swedish Democrats Vänsterpartiet – Left Party Голос(Holos) – Voice/Vote Слуга народу(Sluha narodu) – Servant of the People ВО«Батьківщина»(Batkivshchyna) – All-Ukrainian Union Fatherland Європейська солідарність (Yevropeis'ka solidarnist) – European Solidarity Poland Serbia Serbia Serbia Serbia Serbia Serbia Spain Spain Spain Spain Spain Spain Sweden Sweden Sweden Sweden Sweden Sweden Ukraine Ukraine Ukraine Ukraine 83

How democratically elected politicians get away with autocratizing their country LITERATURE Adorno, T. et al. 1950. The Authoritarian Personality. Harpers. Bonikowski, B. 2017. Ethno-nationalist populism and the mobilization of collective resentment, The British Journal of Sociology 68, 1: 181-213. Brieger, S. A. 2019. Social Identity and Environmental Concern: The Importance of Contextual Effects. Environment and Behavior, 51(7), 828–855. European Election Voter Study. 2019. http://europeanelectionstudies.net/european-election-studies/ees-2019-study/voter-study-2019. Dahl, R. 1971. Polyarchy: Participation and opposition. New Haven, CT: Yale University Press. Dahl, R. 1989. Democracy and its critics. New Haven: Yale University Press. Diamond, L. 1994. Toward Democratic Consolidation, Journal of Democracy 3: 4-17. Graham, M. and Svolik, M. 2020. Democracy in America? Partisanship, Polarization, and the Robustness of Support for Democracy in the United States, American Political Science Review, 114, 2: 392 – 409. Halikiopoulou, D.,& Vlandas, T.(2022). Understanding right-wing populism and what to do about it. Friedrich-Ebert-Stiftung – Democracy of the Future. Vienna Halikiopoulou, D.,& Vlandas, T.(2020). When economic and cultural interests align: the anti-immigration voter coalitions driving far right party success in Europe. European Political Science Review, 12(4), 427-448. Linz, J. and Stepan, A. 1996. Toward Consolidated Democracies, Journal of Democracy 7: 14-33. Levitsky, S. and Ziblatt, D. 2018. How Democracies die. New York: Crown. Lipset, S. M. 1959. Some Social Requisites of Democracy: Economic Development and Political Legitimacy. American Political Science Review 53, 1: 69–105. Mayda, A.M.(2006),‘Who is against immigration? A cross-country investigation of individual attitudes toward immigrants,’ The Review of Economics and Statistics 88(3): 510–530. McCoy, J., Rahman, T. and Somer, M. 2018. Polarization and the Global Crisis of Democracy: Common Patterns, Dynamics, and Pernicious Consequences for Democratic Polities, American Behavioral Scientist 62, 1: 16–42. Pew Researcher Center. 2009. The Post-Communist Generation in the Former Eastern Bloc. https://www.pewresearch.org/global/2010/01/20/the-post-communist-generation-in-the-former-eastern-bloc/ Przeworski, A. et al. 2012. Democracy and Development: Political Institutions and Well-Being in the World, 1950–1990. Cambridge University Press. Norris, P. 2011. Democratic Deficit: Critical Citizens Revisited. Cambridge, UK: Cambridge University Press. Rydgren, J.(2008),‘Immigration sceptics, xenophobes or racists? Radical right-wing voting in six West European countries’ European Journal of Political Research 47: 737–765. Stockemer, D., Halikiopoulou, D.,& Vlandas, T.(2021).‘Birds of a feather’? Assessing the prevalence of anti-immigration attitudes among the far right electorate. Journal of Ethnic and Migration Studies, 47(15), 3409-3436. Svolik, M. 2019. Polarization vs. Democracy, Journal of Democracy 30: 20–32. Varieties-of-Democracy. Democracy Report 2022. Autocatisation Changing Nature? Varieties-of-Democracy. V-Dem Dataset Version 12. Liberal Democracy Index. Polarization of Society Index. https://www.v-dem.net/vdemds.html Weingast, B. R. 1997, The Political Foundations of Democracy and the Rule of Law, American Political Science Review 91: 245-63. 84

IDENTITY, PARTISANSHIP, POLARIZATION Imprint FES Regional Office for International Cooperation Democracy of the Future Reichsratsstr. 13/5, A-1010 Vienna Responsible for Content: Johanna Lutz| Director, Democracy of the Future Tel.:+43 1 890 3811 301 @FES_Democracy democracy.fes.de Orders/ Contact: democracy.vienna@fes.de ISBN 978-3-98628-151-9 Commercial use of all media published by the Friedrich-Ebert-Stiftung(FES) is not permitted without the written consent of the FES. The views expressed in this publication are not necessarily those of the Friedrich-Ebert-Stiftung or of the organization for which the author works. 85

democracy.fes.de