Income inequality and populism
Does income inequality explain the success of populist political parties in Europe?
Master thesis Political Science Master of Science Political Economy
Graduate School of Social Sciences (GSSS) Written by Carlo de Cocq
Student id 11153806
Date June 2018
Supervisor Eelco Harteveld MSc. Second reader Dr. Armen Hakhverdian
Abstract
Populist political parties are increasingly popular in most European countries. Political scientists make effort to explain this increased popularity of populism. Some scholars blame cultural factors where others blame economic factors. This thesis tested whether income inequality explains the populist vote. Evidence from 24 European countries shows that income inequality explains the populist vote in certain regions. Immigration has become an important factor in the popularity of populist parties. This thesis will show how immigration contributes to the relation between income inequality and populism.
Table of content
Introduction ...
4Theoretical framework ...
5Data and Methods ...
9Results ...
11Overall effect of income inequality on the populist vote ... 11
Experienced income inequality and the populist vote ... 16
The effect of income inequality on attitudes toward immigration ... 18
Income inequality and immigration: regional differences ... 19
The relation between views on immigration and the populist vote ... 21
Income inequality and left-wing populism ... 23
Conclusion & Discussion ...
26Bibliography ...
29Appendices ...
31Appendix I; list of ESS variables used ... 31
Appendix II; list of populist political parties ... 32
Introduction
Populist political parties have been increasingly dominant in European politics since the 1960’s. Their share of votes in national parliaments has doubled since the 1980’s. The popular populist parties have changed the political landscape in Europe significantly. The Brexit referendum in the United Kingdom and the following announcement to leave the European Union typify the impact of populist movements on the political landscape (Inglehart & Norris, 2016). Populism received a great deal of attention by media, political and social scientist, politicians and the public (Gidron & Bonikowski, 2013). Much of this attention consists of political scientists trying to explain the populist votes; the reasons why people vote for a populist political party. Debate on what explains the populist vote is roughly divided into two camps: some scholars state that cultural factors best explain the populist vote while others state that economic factors are the best explanation. Inglehart and Norris (2016) for example found western societies to be changing towards more progressive values. An increasingly large group of people does not identify with these progressive values and create a cultural backlash by voting for populist parties. This is a result of populist political parties appointing the shift in cultural values and blaming the elites. After considering all relevant factors, Inglehart and Norris (2016) conclude that this cultural backlash is the most important factor that explains the populist vote (Inglehart & Norris, 2016). Oesch (2008) shows that cultural factors, for example the perception of immigration as a threat to identity, are the main explanation for populist support in Europe (Oesch, 2008). Other scholars view economic factors as the main force behind the success of populist parties. Guiso et al (2018) for example see that economic insecurity, fueled by the economic crisis of 2008, is the main explanation for populist support (Guiso et al., 2018). Gidron & Hall (2017) have found that a combination of changing economic factors and cultural anxiety fueled the increase in votes for populist parties. They acknowledge that the cultural anxiety is fueled by changes in economic status. However, the question remains what specific economic factors explain the populist vote (Gidron & Hall, 2017). Research shows that countries with higher income inequality have lower voter turnouts during elections and a higher backlash on globalization (Burgoon, 2013). This thesis aims to solve a part of the populist puzzle by analyzing the economic mechanism of income inequality behind the populist vote. It considers other factors, such as immigration, while analyzing both the perceived income inequality as well as the country-level data. The goal is to answer the following question: ‘Does income inequality explain the success of populist political parties in Europe?’
Theoretical framework
This thesis aims to give an explanation for the populist vote. More specifically, it aims to understand the relationship between income inequality and the vote for populist political parties. In order to do this, a thorough understanding of the concept of populism and the theoretical explanations of populism are necessary. The definition of populism and populist parties depends on the context and shows varieties across different fields of study. However, there is academic consensus that the definition used by one of the leading political scientists on populism, Cas Mudde, is sufficient. Mudde (2016) describes populism as an ideology that divides society between two antagonistic groups; the ‘pure’ people and the ‘corrupt’ elites. Populist political parties aim to shift power away from the ‘corrupt’ elites and towards the pure people. In most European parliaments, parties that can be defined as populist are represented and most of these parties are considered to be right-wing political parties (Mudde, 2016).
The literature identifies a debate between scholars that focus primarily on cultural aspects and scholars that focus on economic aspects as the main explanations for the populist vote (Kriesi et al., 2012; Inglehart and Norris, 2016; Gidron & Hall, 2017). Inglehart and Norris (2016) identify two perspectives on populism and populist political parties. The economic inequality perspective identifies changes in the workforce, education and globalization. It views them as the main reasons why voters, especially in the middle class, shift towards populist political parties. The cultural backlash perspective blames progressive values such as multiculturalism and gender equality. An increasingly large group of people cannot identify with these values, which explains the rise of populism (Inglehart & Norris, 2016).
The arguments on the side of the economic inequality perspective have gathered support in recent decades. There is evidence that societal changes, caused by globalization and the capitalist system itself, are one of the main explanations for the growing popularity of populism (Guiso et al., 2018). Green (2017) for example identifies economic inequality in general and income inequality in particular as the main reasons for dissatisfaction against governments. Populist political parties are channeling this dissatisfaction and using it to receive growing support. Populists do this by blaming the ‘corrupt’ elites for policies that caused economic inequality to increase (Green, 2017). Economic uncertainty makes more voters vulnerable for populist appeal. As income inequality increases, these voters will turn away from the incumbent parties and vote increasingly more for populist parties (Aggeborn, 2017). Research by Winkler (2014) shows that rising income inequality leads to more political polarization and, especially among older middle-class voters, an increase in votes for populist political parties (Winkler, 2014). Oesch (2008) explains that voters most affected by globalization and changes in the
workforce and technology make up most of the group that support populist political parties. Evidence from four Western-European countries shows that workers are most prone to populist appeal because ‘unlike qualified employees who benefit from technological progress and the opening of borders, workers often lack convertible skills necessary to adjust to these new circumstances’ (Oesch, 2008, p. 351). The arguments mentioned by Guiso et al. (2018), Green (2017), Winkler (2014) and Oesch (2008) suggest that an increasing group of voters, often workers, see their relative economic position decrease and are increasingly exposed to economic insecurity. Kriesi et al. (2012) call this group of voters the ‘losers of globalization’ and see societies changing towards new cleavages that puts these ‘losers and winners of globalization’ opposite each other (Kriesi et al., 2012). In the literature, this ‘winners versus losers of globalization’ theory is often used as explanation for the rise of populism (for example: Cuperus, 2017; Rooduijn, 2015; Kriesi, 2014; Kriesi et al., 2012). This ‘winners versus losers of globalization’ theory however, consists of more factors than mere economic insecurity or economic globalization. Literature also identifies cultural factors that are in play when explaining the populist vote.
The cultural backlash perspective in the literature focuses on changing values in western societies as well as perceived threats to the collective identity of the supporters of populism. The cultural backlash theory states that the rise of populism can be explained by the fact that an increasingly large group of people reject the modern-day, progressive values that are dominant in western societies (Inglehart & Norris, 2016). Issues such as climate change, gender equality, gay-rights and positive attitudes toward immigrants are threatening the values of the more traditional voters. These voters tend to turn toward populist political parties, especially since these parties blame the elites for pushing and promoting said values. Populist political parties can also frame these issues as threats to national identities, or the identities of the ‘pure’ people (Guiso et al., 2017). Some scholars regard these cultural factors as the most important explanation of the shift towards populist political parties (for example: Kriesi et.al, 2012; Oesch, 2008; Inglehart & Norris, 2016). Some of these cultural factors, however, directly intertwine with the economic perspective on populism. Immigration for example can be viewed as threatening to the identity of a group or their cultural values. It can also be perceived as a purely economic threat; the fear to lose jobs to people from other countries. Moreover, the group that is most prone to the anti-immigration appeals of populist parties is also known to be the group that is threatened by immigration in an economic way (Guiso et al., 2017).
The two sides of the debate both provide useful explanations for the increasing number of votes for populist parties. Choosing either side of the debate, however, will not provide an answer to the question posed in this thesis. The two perspectives are intertwined on different
levels and in multiple ways. Research shows that the core supporters of populist political parties consists of working class people that perceive their social status, income and cultural values as declining (Gidron & Hall, 2017). Income inequality and the perception of it can contribute to two mechanisms in play. First of all, income inequality can create a wider gap between groups in society as well as make the group of dissatisfied middle-class voters larger (Aggeborn, 2017). Secondly, the perceived income inequality can create a situation where this group feels increasingly deprived compared to others in society (Burgoon, 2013; Aggeborn, 2017). Finally, the social identity theory indicates that in situations with high income inequality, lower class voters switch their focus from economic to cultural issues. This means lower-class voters find it harder to identify with other ‘poor’ voters and identify more with others on cultural issues (Han, 2016). These factors stimulate and increase the cultural backlash in a way that the group that feels economically deprived, although relative to others in the same society, is more susceptible for the cultural appeal of populist parties (Gidron & Hall, 2017).
Two perspectives on the increasing success of populist parties have been discussed. The immigration-mechanism that connects the two perspectives has been explained as well. An analysis of the literature so far has identified that income inequality could fuel negative views on immigration, which can instigate populist support. Han (2016), Winkler (2014), Green (2017), Nolan (2017) and Aggeborn (2017) found a direct influence of income inequality on populism. This indicates that income inequality could be the key in explaining the populist vote. The first hypothesis reflects that direct effect between income inequality and the populist vote.
H1: when income inequality rises in a country, more individuals vote for a populist party People vote based on a number of factors that influence their policy preferences. Theories of economic voting do state that the incumbent party or parties can be punished for the state of the national economy (See for example: Hellwig, 2012; Lewis-Beck & Nadeau, 2011). Research conducted in the United States, however, shows that this theory only works when the specific economic situation of different groups is considered. Voters could be unaware of the macro-economic factors or simply not experience a change in these conditions (Linn & Nagler, 2017). Testing the income inequality of a country against the share of populist votes could encounter this problem. Therefore, it is important to test the experienced income inequality, or specifically the people that regard income inequality as a problem and feel that something should be done about it.
H2: when income inequality is recognized as a problem, individuals are more likely to vote for a populist party
The literature identifies attitudes toward immigration as an influential factor for explaining the populist vote. Hence, it is important to test the effect of income inequality on attitudes toward immigration and the effect of these attitudes on the populist vote. The impact of immigration-attitudes on electoral results has been tested before. Research indicates that more negative views on immigration are related to more votes for populist political parties (see for example; Lucassen & Lubbers, 2012; Rooduijn et al., 2017). The impact of income inequality on these views however has not yet been tested directly.
H3.1: when income inequality rises, views on immigration become more negative
The effect negative views on immigration have on populist political parties, which has been tested before, connects the pieces of the puzzle. By showing the mediating effect of immigration between income inequality and the populist vote, the populist vote could be explained.
H3.2: the more negative the attitudes toward immigration, the more individuals that vote for a populist political party
Immigration could be an important factor in explaining the support for populist parties. However, immigration remains a typical issue that distinguishes the populist right from the populist left (Lucassen and Lubbers, 2012). Predictions of this thesis are that for populism in general and right-wing populism in specific, income inequality has an indirect effect on electoral results. However, literature indicates that left-wing populism cannot be explained through immigration (Rooduijn et al., 2017). The demand-side for left-wing populism consists of voters that are disappointed by elites and blame them for economic insecurity and income inequality (Guiso et al., 2017). Winkler states that income inequality, because of the sentiments described above, fuels the support for left-wing populism (Winkler, 2014). Therefore, a direct effect of income inequality on left-wing populism is to be expected.
H4: the higher the income inequality, the more likely an individual is to vote for a left-wing populist party
Together, the evidence gathered by testing the hypotheses will answer the question whether or not income inequality can explain the populist vote. This chapter described the theoretical framework necessary for answering the research question. The next chapter will describe how the posed hypotheses will be tested and what data is going to be used.
Data and Methods
The cases selected to answer the research question come from the European Social Survey database. The European Social Survey (ESS) is conducted periodically, with a new survey round every two years. It is used by many scholars for research on populism (For example: Algon et al., 2017; Guiso et al., 2017; Han, 2016; Winkler, 2014). Data from all available ESS-rounds regarding two topics; politics and socio demographics, are merged and used as the base of the dataset. The dataset will then be adjusted in order to contain only the variables relevant for this research; variables regarding the political party voted for in the last election and variables regarding views on immigration and perceived income inequality. Appendix I shows the names of the ESS variables that the dataset uses and the corresponding questions in the European Social Survey rounds. In order to select populist political parties in the database, research by Inglehart and Norris (2016) is used as guidance. For their research, Inglehart and Norris used the 2014 Chapel Hill expert survey to indicate which parties in Europe can be defined as populist parties. The Chapel Hill expert survey uses expert opinion to obtain the ideological position of political parties (Bakker et al., 2012). To create a variable that indicates whether or not a political party can be regarded as populist, a dummy variable has to be created, where ‘1’ represents populist parties and ‘0’ represents other political parties. In this thesis, the Chapel Hill dataset itself is used to identify parties as populist left or populist right, based on their overall ideological position. Inglehart and Norris on the other hand, defined parties as left or right based on their economic positions. The reason for using the overall ideological position instead of the economic position is that the position of many populist parties can be regarded as economic left. However, populist parties distinguish themselves more on other issues such as immigration, in which they can be regarded as right-wing (Harteveld, 2016). A complete overview of all parties that are labeled as populist and their ideological orientation can be found in appendix II.
The dataset contains around 300.000 observations from 24 countries. These countries are selected because they represent the greater part of the European Union and vary from Northern and Western to Southern and Eastern-European countries. Some countries of the European Union are unavailable in the European Social Survey data and will therefore be missing from the dataset. Latvia, Malta and Romania are missing in the European Social Survey data and therefore left out of the dataset. Using a variety of countries is necessary to measure the effect of income inequality on populist votes in different countries. However, this does not mean that the effect is measured between countries or that the countries are comparable. The effect is measured within countries and over time. This means a change of the GINI coefficient
would result in a change in, for example, the number of individuals that vote for a populist party. In order to test the effect on different sets of countries, different models are created for the specific regions. Accordingly, variables are created that contain the data from the countries within this region only, regarding the different topics. In addition, a region variable is going to be created that divides the countries in the dataset into five different regions; North, West and South, North and West, East, North and Scandinavia. These countries are sorted into these regions according to the definition of the United Nations. The list of regions can be found in appendix III. This division is essential in order to properly use the dummies for the country fixed effects that are included in the regression analysis.
GINI data has been retrieved from the World Bank online database for the years 2002 to 2014. The dataset containing the GINI data for all countries and all years is merged with the database containing the European Social Survey answers. Literature indicates there is a possibility of the GINI coefficient not immediately affecting individuals’ views or voting behavior. To test the effect of income inequality on the populist vote, experienced income inequality must be considered. The European Social Survey does not contain questions that directly ask individuals whether or not they experience income inequality. The ESS-variable ‘gincdif’ contains the answers to the question whether or not ‘governments should reduce differences in income levels’. Individuals that (strongly) agree with this statement identify income inequality as a problem and feel that something should be done about it. The variable is renamed ‘perceived inequality’ since it is used in this dataset to test the perceived income inequality. The variable has a shortcoming that should be taken into consideration. The variable does not contain answers to a question that asks respondents whether or not they experience income inequality directly. It asks whether or not something should be done about inequality, regardless of the respondent’s personal situation. This could imply that an individual with a relatively high and secure economic position could think income inequality should be addressed. The ‘perceived inequality’ variable is the only variable in the ESS that relates to income inequality. Relative to the GINI coefficient it is a more reliable variable to operationalize perceived inequality, since it asks if individuals view income inequality as a problem.
Views on immigration are measured with a new variable that combines different variables from the European Social Survey. The first variable, ‘imbgeco’ asks respondents if immigration is good or bad for the country’s economy. The second variable, ‘imueclt’, asks whether the country’s cultural life is enriched or undermined by immigrants. The third and last variable, ‘imwbcnt’ asks whether immigrants make a country a worse or a better place to live.
The variables are tested for internal consistency which resulted in a reliability coefficient of 0.846 (Cronbach’s α = 0.846).
The relations between the variables are tested through a logistic regression with country fixed effects. By using a logistic regression with dummies for the country fixed effect it is possible to analyze the differences within countries, over time. Because the macro-level GINI data is merged with the individual level ESS data, each individual obtains its ‘own’ GINI observation. In order to prevent an overload of GINI observations during the regression, the outcomes are clustered in waves of country-year (for example; Austria, 2002).
Results
The aim of this thesis is to explain the relationship between income inequality and the vote for populist political parties. The different hypotheses will be tested and briefly discussed in this section. First, the overall effect of income inequality on the populist vote is tested. Then, this effect will be tested for specific regions. Third, the effect of perceived income inequality on the populist vote is tested. This effect will be tested in general as well for specific regions. The effect of income inequality on immigration and that of immigration on the populist vote will be tested, first for the complete dataset and then again for the different areas. Finally, left-wing populism will be further analyzed by testing the relation between income inequality and votes for left-wing populist political parties.
Overall effect of income inequality on the populist vote
Literature indicates that rising income inequality could have a correlation with electoral successes of populist political parties. Testing the relation directly, however, shows no significant correlation between the GINI coefficient and the vote for a populist party. Table 1 shows the outcome of a logistic regression for the GINI variable and the variable that
indicates whether a person has voted for a populist political party. Since p > 0.05 there is no convincing evidence that income inequality, in the form of the GINI coefficient, has a significant relation with votes for populist parties.
The results shown above are explainable by the fact that individuals might not directly experience changes in the GINI coefficient. Marco-economic conditions in a country are not always perceived by voters (Linn & Nagler, 2017). The GINI coefficient, which measures income inequality, is a macro-economic factor and is therefore not directly experienced by voters. Scholars indicated that macro-economic factors such as income inequality could directly affect the support for populism (Han, 2016; Green, 2017). The results so far do not provide any evidence pointing in this direction. Despite the fact that the results so far show no evidence that prove the relationship between income inequality and the populist vote, further tests can still contribute to the puzzle. The next sections will test the effect of regional differences on the relation between income inequality and the populist vote. Following, perceived income inequality will be taken into account.
(1)
Variables Populist vote
Gini 0.0615 (0.579) (0.185) Constant -3.893 (0.249) Observations 82,541 Pseudo R² 0.1887 Country fixed effects included
Robust pval in parentheses ** p<0.01, * p<0.05, ⁺ p<0.1
It can be concluded that there is no direct effect of income inequality on votes for populist parties. Nevertheless, these tests were conducted using data from all 24 countries. The set of countries consists of Northern-, Western-, Southern- and Eastern-European countries. The first tests measured the effect of income inequality on the populist votes within these countries and over time, but the overall effect remained insignificant. Since scholars used different country-samples, the mechanism could still apply to specific regions. Green (2017) for example used eleven Western democracies and Oesch (2008) five different countries. To analyze whether the mechanism still applies to specific regions, the effect of income inequality on votes for populist parties was tested in different regions. The regions are defined in appendix III and based on the definition provided by the United Nations. The mechanism was tested in Northern-, Western and Southern-European countries combined. The results in table 2 shows the results of this test. The outcomes of the regression analysis are insignificant since p > 0.05. The test for Eastern-European countries show similar results; an insignificant effect of income inequality on whether or not individuals voted for a populist political party.
Table 2; North, West and South Table 3; Eastern Europe
Table 3 shows the regression results of the test for the Eastern-European region and establishes that the effect is not significant. The division between Eastern-European and other countries in
(1) Variables Populist vote
North, West & South Gini North, West &
South -0.0971 (0.146) Constant -0.0994 (0.960) Observations 122,473 Pseudo R² 0.1293
Country fixed effects included Robust pval in parentheses ** p<0.01, * p<0.05, ⁺p<0.1 (1) Variables Populist vote East Gini East -0.0920 (0.651) Constant 1.368 (0.844) Observations 39,045 Pseudo R² 0.1120 Country fixed effects included
Robust pval in parentheses ** p<0.01, * p<0.05, ⁺p<0.1
Europe does not provide any convincing evidence that the effect of income inequality on populist votes is significant for these regions.
Tests with various smaller regions resulted in similar outcomes. Table 4 shows that there is not a significant effect for income inequality on populist votes; not in either Northern and Western countries combined or in Northern-European countries alone. For these regions, results of the test show that p > 0.05. Testing the income inequality against votes in Scandinavia, however, does show a significant positive effect (r=.13, p < 0.01). Consequently, in Scandinavia a higher income inequality predicts more votes on populist political parties.
Table 4; North/West, North, Scandinavia
Scholars on the economic side of the debate determined income inequality to be an explanation for the increasing popularity of populist political parties. Green (2017) indicated that rising inequality could increase the dissatisfaction against governments, fueling the populist parties’ success (Green, 2017). According to the ‘winners versus losers’ theory, the ‘losers of globalization’ see their relative economic position decline and they form the largest part of the populist parties’ supporters (Kriesi et al., 2012). Burgoon (2013) stated that economic inequality and income inequality could be the economic factors that instigate dissatisfaction and support for populism (Burgoon, 2013). Aggeborn (2017) stated that income inequality moves more voters away from traditional parties towards populist political parties (Aggeborn, 2017). Results found in this thesis, however, show that income inequality does not
(1) (2) (3)
Variables Populist vote North
& West
Populist vote North Populist vote Scandinavia Gini North & West -0.0426
(0.547) Gini North -0.0314 (0.755) Gini Scandinavia 0.134** (0.00872) Constant -1.732 -1.664 -6.090** (0.410) (0.534) (1.14e-05) Observations 104,683 46,081 20,181 Pseudo R² 0.0639 0.0671 0.0640 Country fixed effects included
Robust pval in parentheses ** p<0.01, * p<0.05, ⁺p<0.1
directly correlate with more support for populist parties. The only exception is Scandinavia, where rising income inequality leads to more support for populist parties. Graph 1 shows that higher income inequality increases the chance of people voting for populist parties in Scandinavia. When the GINI-coefficient increases from 30 to 40 the chance of voting for a populist party doubles. These results are remarkable considering all GINI-coefficient observations from Scandinavian countries are on the far low side of the spectrum. Since the change of the Gini is tested within countries and over time, an explanation could be that Scandinavian countries experienced a relatively high change in income inequality in a relative short period of time. Regardless, this would not unequivocally explain the strong effect found in the regression analysis. As Linn and Nagler (2017) suggested, the GINI coefficient does not have to be ‘experienced’ by voters in a country. So far, the conclusion that can be drawn with convincing evidence is that hypothesis 1, suggesting the relation between income inequality and populist votes, can be rejected for all regions except for Scandinavia.
Experienced income inequality and the populist vote
The first hypothesis is rejected for all regions except Scandinavia. This was to be expected, since the literature suggested that voters might not experience macro-economic indicators such as income inequality. Therefore, experienced income inequality should also be tested. Testing experienced income inequality against whether or not individuals voted for a populist party shows an interesting relation between the two variables. The variable ‘perceivedinequality’ contains the respondent’s answer to the question whether or not the government should do something about income inequality. Table 5 shows the outcome of the regression that tested if individuals are more or less likely to vote for a populist party when in agreement with the statement that the government should do something about income inequality. The regression analysis resulted in an effect that is not significant within
the 95 percent confidence level, since p > 0.05. Therefore, it does not provide enough evidence to accept hypothesis 2. This outcome requires more consideration. The Data & Methods section already discussed some of these considerations, but this outcome demands some more remarks. The variable used in this regression tests the relation between attitudes toward reducing income inequality and the vote for populist party. This means individuals with a relatively well-of economic position might still think differences in income should be reduced. The variable does not directly state that an individual perceives the economic situation in a country as ‘unequal’. In conclusion, the outcomes shown in table 5 do not provide more than some suggestive evidence.
It has been concluded that a significant effect was found for income inequality on the populist vote in the Scandinavian region. Therefore, it is useful to compare the different regions for perceived inequality as well. The first
test between experienced income inequality and the populist vote only provided results regarding the entire dataset containing observations from 24 countries. The second test will compare specific European regions. A division is made between Northern- and Western-European countries, Northern-Western-European countries and Scandinavia. While income inequality measured by the GINI coefficient showed significant results for Scandinavia, perceived income
(1) Variables Populist vote Perceived inequality -0.0418⁺ (0.0822) Constant -2.688** (0) Observations Pseudo R²
Country fixed effects included
153,190 0.1607
Robust pval in parentheses ** p<0.01, * p<0.05, ⁺p<0.1
inequality on the other hand does not. Testing the perceived income inequality against the populist vote does not bring forward any convincing evidence. No significant effect was found in the test that contained all regions and no significant effect was found for the specific regions. Table 6 provides the results of the regression analysis conducted in the region-specific model. For all regions, p > 0.05. Hence, no evidence is found that there is a relation between perceived inequality and votes for populist parties. Hypotheses 2 stated that the more people perceiving income inequality as a problem, the higher the number of votes for populist parties. Hypothesis 2 is rejected.
(1) (2) (3)
Variables Populist vote North
& West Populist vote North Populist vote Scandinavia
Perceived inequality North & West
-0.00640
(0.832)
Perceived inequality North -0.0663
(0.158) Perceived inequality Scandinavia 0.0350 (0.173) Constant -3.050** -2.315** -2.615** (0) (0) (0) Observations 134,016 52,538 23,041 Pseudo R² 0.0714 0.0744 . 0.0712
Country fixed effects included Robust pval in parentheses ** p<0.01, * p<0.05, ⁺p<0.1
The effect of income inequality on attitudes toward immigration
Hypothesis 1 is now partly rejected. Only for Scandinavia the effect of income inequality on votes for populist political parties turned out to be significant. The second hypothesis regarding perceived inequality is rejected completely. However, this does not mean the literature does not provide any more clues for explaining the populist vote. Some of the mediating effects should still be tested based on the indications given by the literature. The debate on the role and influence of immigration on the populist vote is not yet settled. Nonetheless, many scholars mention that attitudes toward immigration matter in the sense that they have to be considered in explaining the populist vote. Kriesi et al. (2012) for instance found that attitudes toward immigration are one of the main issues that dominate the support for populist political parties (Kriesi et al., 2012). Winkler indicated that income inequality increased the anti-immigration sentiment (Winkler, 2014). The relation between income inequality and immigration will be tested in two steps. First, the effect of income inequality on views on immigration is tested. Then, the effect of immigration views on the populist vote is tested. The first test clearly indicates that when the income inequality, represented by the GINI coefficient, rises, the views on immigration become more negative. This effect is significant (p < 0.05) and shows a strong negative correlation (r=-.11) between income inequality and views on immigration, as can be seen in table 7. (1) Variables Immigration Gini -0.114* (0.0490) Constant 6.354** (0.000304) Observations 146,921 Pseudo R² 0.0560 Country fixed effects included
Robust pval in parentheses ** p<0.01, * p<0.05, ⁺p<0.1
Table 7; income inequality views on immigration
Hypotheses 3.1 is accepted based on the evidence shown in table 7. This suggests that the economic inequality perspective on populism is useful in the sense that rising inequality creates a larger group of people that become critical on immigration. The ‘immigration’ variable consists of three highly correlated variables that include views on immigration in an economic
and cultural way. Theory suggested that immigration is an issue that intertwines cultural and economic factors; immigration threatens both the economic and cultural position of some individuals. The two perspectives on explaining the populist votes clearly come together here. Even though the overall correlation between income inequality and populist votes is insignificant, there is still a strong connection between income inequality and immigration that needs explanation. One explanation is the theory that suggests economic inequality creates a larger group of people that becomes susceptible for anti-immigration views (Burgoon, 2013; Aggeborn, 2017; Guiso et al., 2017). Another example of this mechanism is the social identity theory which states that in situations of higher income inequality, voters focus on issues that threaten their social identity (Han, 2016). The results of this analysis show that income inequality and views on immigration are correlated.
Income inequality and immigration: regional differences
The only evidence of a correlation between income inequality and populist votes came from Scandinavia. Analysis of Scandinavian countries was the reason hypothesis 1 could not be fully rejected. These results require more explanation, since it is not clear why rising inequality causes increasing support for populist parties in that region. Immigration could be the explaining factor and the effect of income inequality on views on immigration is therefore tested for specific regions.
(4) (5)
Variables Immigration North Immigration Scandinavia
Gini North 0.0426 (0.481) Gini Scandinavia 0.126 (0.229) Constant 3.437* 1.230 (0.0362) (0.665) Observations 59,162 20,062 Pseudo R² 0.0498 0.0086 Country fixed effects included
Robust pval in parentheses ** p<0.01, * p<0.05, ⁺p<0.1 Table 8; North and Scandinavia
Table 8 shows that no significant effect is found for the Northern-European countries. For the region defined as ‘North’, p > 0.05 and therefore the results are not significant. For Scandinavia, the p > 0.05 as well and thus no significant relation can be found.
It is remarkable to observe that the regression analysis for Scandinavia did not result in a significant correlation between income inequality and views on immigration. This is because the region did show a significant positive correlation between income inequality and votes for populist parties. Since the strongest correlation cannot be found in Scandinavia or Northern-Europe, it has to be tested whether other regions can explain the positive effect. By testing the effect for the other regions, the results can be compared accordingly. Table 9 shows that a significant correlation can be found for the Eastern-European countries (p < 0.01). This means the effect comes from countries in the Eastern-European region. There, the views on immigration become more negative when the income inequality rises. The effect of income inequality on the views on immigration is insignificant for all other regions.
(1) (2) (3)
Variables Immigration North, West & South
Immigration East Immigration North & West
Gini North, West & South
-0.0776 (0.136)
Gini East -0.148**
(0.000717)
Gini North & West 0.0214
(0.524)
Constant 5.477** 8.552** 2.668**
(0.000485) (3.23e-08) (0.00743)
Observations 138,825 37,285 123,030 Pseudo R² 0.0625 0.0238 0.0373 Country fixed effects included
Robust pval in parentheses *** p<0.01, ** p<0.05, * p<0.1
Table 9; regional differences income inequality and views on immigration
Where other mechanism proved to be significantly correlated in Scandinavia, the relation between income inequality and views on immigration is not. For the Northern and Scandinavian
region, correlations are insignificant. Hypothesis 3.1 is accepted and the effect comes from Eastern-European countries.
The relation between views on immigration and the populist vote
The second part of proving the relationship between income inequality and the populist vote through immigration is testing the effect of immigration views on populist support. The effect of views on immigration on the populist vote is a highly significant (p < 0.01), negative effect (r=-.165). Thus, the more positive an individual becomes about immigration, the less likely he or she is to vote for a populist party. Hypotheses 3.2 can be accepted based on the evidence provided by table 10. Oesch (2008) already established that cultural grievances over immigration largely explain the populist vote in five European countries. Guiso et al. (2017) projected that the increased popularity of populism has an economic cause through a cultural mechanism such as attitudes toward immigration (Guiso et al., 2017). In sum, income inequality fuels the negativity toward immigration and that explains the rise of populist votes.
Table 10; immigration populist vote
The effect is tested for the different regions since earlier results showed differences between the specific regions. Table 11 shows that the correlation is negative for all regions. So, more positive views on immigration lower the chance of voting for a populist political party. With the exception of Eastern-European countries, the relation is significant. The effect in the region North, West and South is highly significant with (p < 0.01). Similar results are found in the region North & West and in the Northern region (p < 0.01). The strongest effect (r= -.510) can be found in Scandinavia and this effect is highly significant because (p < 0.01).
(1) Variables Popvoted Immigration -0.165** (0) Constant -2.025** (0) Observations 153,488 Pseudo R² 0.1750 Country fixed effects included
Robust pval in parentheses
The effect becomes stronger in the more Northern-European countries, with Scandinavia having the strongest negative correlation. This is the opposite effect of income inequality on attitudes toward immigration, where only the Eastern-European region showed a significant correlation. For the Northern, Southern and Western countries, views on immigration are highly correlated with voting for populist political parties.
(1) (2) (3) (4) (5) Variables Populist vote North, West and South Populist
vote, East vote, North Populist and West
Populist
vote, North Populist vote, Scandinavi
a Immigration North,
West and South
-0.283** (0)
Immigration East -0.0260
(0.195) Immigration North &
West -0.378** (0) Immigration North -0.428** (0) Immigration Scandinavia -0.510** (0) Constant -1.599** -1.678* -1.524** -0.449** -0.118 (6.38e-07) (0.0141) (3.91e-10) (0.00644) (0.402) Observations 106,617 40,611 135,519 53,293 23,510 Pseudo R² 0.1467 0.1093 0.1340 0.1541 0.1936 Country fixed effects included
Robust pval in parentheses ** p<0.01, * p<0.05, ⁺p<0.1
Table 11; regional differences immigration populist vote
Hypothesis 3.2 is accepted based on the evidence provided. There is a significant negative effect between attitudes toward immigration and voting for populist parties. Together with the acceptance of hypothesis 1, it is concluded that there is a significant relation between income inequality and immigration-views and between immigration views and votes for populist parties. These findings correspond with the indications given by the literature. Theory stated
that negative views on immigration are related to votes for populist parties (Lucassen & Lubbers, 2012). Attitudes towards immigration are the best explanatory mechanism for explaining right-wing populism (Rooduijn et al., 2017).
Income inequality and left-wing populism
Literature indicated that the populist left distinguishes from the populist right by focusing on different issues. Income inequality could increase the support for left-wing populist political parties (Winkler, 2014). Where the populist right focusses more on immigration and other threats to cultural identity, the left focusses more on inequality (Guiso et al., 2017). The views on immigration by left-wing and right-wing populist supporters are almost opposite (Rooduijn et al., 2017). Supporters of populist political parties in general blame the elites for their situation, left-wing and right-wing alike (Rooduijn et al., 2017). Supporters of left-wing populism blame the elites for a declining economic situation, while focussing more on the government’s inability to change economic inequality (Guiso et al., 2017). Testing the correlation between income inequality and left-wing populism results in a noteworthy effect. Table 12 contains the outcome of the regression analysis that was conducted. The results indicate that rising income inequality results in increased support for left-wing populist parties. The correlation found is positive (r=.130) and significant (p < 0.05), which means a rising income inequality would predict an increase in the number of votes for left-wing populist parties. (1) VARIABLES Populistleft Gini 0.130* (0.0131) Constant -7.420** (0.000120) Observations 19,033 Pseudo R² 0.0391 Country fixed effects included
Robust pval in parentheses ** p<0.01, * p<0.05, ⁺p<0.1
Table 12; income inequality, left-wing populist vote
Based on the results displayed in table 12, hypothesis 4 is accepted. There is convincing evidence that rising income inequality leads to more votes for left-wing populist parties. Graph
2 shows the predictive margins. The chance of an individual voting for a left-wing populist party increases when the GINI coefficient increases. Although Guiso et al. (2018) suggested that economic insecurity is an important factor on the demand-side of left-wing populism, the correlation was not directly tested by others before. Rooduijn et al. (2017) stated that the distinguishing factor between populist left and right is income inequality. Hypothesis 4 is accepted and evidence is provided that income inequality and votes for the populist left are indeed correlated.
The overall test of the relation between income inequality and votes turned out to be insignificant. The exception there is Scandinavia and so it is important to test in what way the effect of income inequality on left-wing populism varies per region. Table 13 shows the correlation becomes weaker when Southern-European countries are no longer included in the model. This implies the effect can mainly be found in the Southern-European region. Some scholars already suggested that left-wing populism is most dominant in Southern-European countries (Rooduijn et al., 2017). The results of the regression analysis prove that there is a strong, significant correlation between income inequality and the vote for left-wing populist parties. Furthermore, it shows that this effect is the strongest when Southern-European countries are included in the model. In contrast to Scandinavian countries, the GINI values of Southern-European countries are all in the high end of the spectrum. The trend is that left-wing populism is on the rise in all of Southern-Europe, except for Italy (Boros et al., 2016). Greece and Portugal have high GINI-scores and score high for anti-globalization sentiments (Burgoon, 2013). The economic crisis left Spain and Greece with a sustained rise of income inequality (Nolan, 2017). These factors all explain why Southern-European countries experience the strongest correlation between income inequality and the left-wing populist vote.
This section showed the results of the different tests that were conducted in order to answer the research question. The tables and graphs containing the regression results were accompanied by short theoretical feedback. The next section will provide an overview of the conclusions based on the research conducted in this thesis. While providing more theoretical reflection, it will answer the research question.
(1) (2) Variables Left-wing populism North, West & South Left-wing populism North & West Gini North, West and South 0.130* (0.0131) Gini North &
West 0.0878* (0.0351) Constant -7.420** -4.843** (0.000120) (1.58e-05) Observations 19,033 9,336 Pseudo R² 0.0391 0.0018 Country fixed effects included
Robust pval in parentheses ** p<0.01, * p<0.05, ⁺p<0.1
Conclusion & Discussion
The goal of this thesis was to answers the following question: ‘Does income inequality explain the success of populist political parties in Europe?’. Research was conducted using ESS-data from 24 European countries in order to find the answer to this question. This section will provide the answer to the research question and an overview of which hypotheses were accepted and rejected. Moreover, it provides a theoretical interpretation of the results as well as implications for the introduced literature.
Income inequality does not explain the success of populist political parties in Europe. It does explain the rise of populism in Scandinavia and left-wing populism. Literature indicated that income inequality could explain the increasing popularity of populist political parties. Using a dataset containing over 300,000 observations, this effect was tested with a logistic regression. No significant effect was found between GINI-data and votes for populist parties. This thesis did not find any evidence for the statements of Han (2016), Winkler (2014), Green (2017) and Aggeborn (2017). They claimed that populism is fueled directly by increasing income inequality. The literature showed that scholars tested their arguments using different sets of countries. The correlation between income inequality and the populist vote was tested using different sets of countries, to see whether there were regional differences. This test showed that a significant positive correlation was found for Scandinavia. Thus, rising income inequality leads to more votes for populist political parties in that area. The GINI coefficients for Scandinavian countries are all on the low end of the spectrum. Income inequality has been rising in Scandinavia in recent years and is now approaching the European average. For Scandinavian countries, GINI levels near the European average are regarded as high (Nolan, 2017). Traditional socio-economic cleavages in Scandinavia have become less significant. Populist parties in Denmark and Sweden have filled the void by opposing economic globalism and cosmopolitism (Rydgren, 2014). The ‘winners versus losers of globalization’ theory applies here. Together with relatively fast-growing income inequality, they explain the populist vote in Scandinavia.
Since macro-economic factors are not always experienced by voters (Linn & Nagler, 2017), the correlation between experienced income inequality was tested. No evidence was found to prove experienced income inequality is correlated with the populist vote. Literature showed that immigration could be an explanation for the populist vote. This thesis provides evidence that attitudes towards immigration are indeed an explanation for the populist vote. Rising income inequality is correlated with negative views on immigration, which means there is an indirect effect between income inequality and the populist vote through immigration.
Since immigration is a typical right-wing issue, it cannot explain the votes for left-wing populist parties. Following the literature, the effect of income inequality on the populist vote was tested and proved to be significant. In conclusion, increasing income inequality results in more votes for populist left-wing parties.
Hypotheses 1 stated that increasing income inequality would result in more votes for populist parties cannot be accepted since there is no convincing evidence for this statement. The only exception is the Scandinavian region. Scholars on economic voting indicated that voters do not directly ‘experience’ macro-economic factors. This problem was encountered when the GINI-coefficient was used to measure income inequality in this thesis, since the GINI GINI-coefficient is a macro-economic indicator. The effect of experienced income inequality on the populist vote was tested and no significant effect was found. All things considered, there is no evidence that supports hypothesis 2, which stated that ‘when income inequality is recognized as a problem, individuals are more likely to vote for a populist party’.
The debate on explaining the populist vote is roughly divided into the economic inequality perspective and the cultural backlash perspective. Immigration is an issue that is reflected on both sides of the debate. The correlation between income inequality and views on immigration was tested as well as between views on immigration and the populist vote. Views on immigration become more negative when the income inequality increases. Results of the conducted test provide evidence for the idea that rising income inequality could fuel cultural backlash. This is what Burgoon (2013), Han (2016) and Guiso et al. (2017) suggested with regards to immigration specifically.
The effect found in this thesis regarding immigration comes from Eastern-Europe primarily. No significant effect was found for the Northern, Southern, Western and Scandinavian regions. Hypothesis 3.1, which stated that ‘when income inequality rises, views on immigration become more negative’ is accepted with the remark that the effect comes from Eastern-European countries.
When views on immigration become increasingly negative more people vote for a populist political party. Results of the conducted logistic regression indicate that when views on immigration become more positive, less people vote for populist parties. Opposite to the findings of the relation between income inequality and immigration, the effect was significant in every region except for Eastern-Europe. The effect was found to be strongest in Scandinavia.
Hypothesis 3.2 which stated: ‘the more negative the attitudes toward immigration, the more individuals that vote for a populist political party’ is accepted.
Views on immigration are an important factor in explaining the increasing number of votes for populist parties. Immigration however is an issue that distinguishes the populist right from the populist left. In explaining the vote for left-wing populist political parties, literature indicated that there could be a direct relation with income inequality. A significant positive effect was indeed found after testing income inequality and the left-wing populist vote. This effect comes from the Southern-European countries, which have experienced an increase in the number of votes for left-wing populist parties (Boros et al., 2016). This breaks with the trend elsewhere in Europe, where right-wing populist parties are more successful. Italy is the only exception in Southern-Europe, as right-wing populist parties are increasingly successful there (Boros et al., 2016). The significant effect for Southern-Europe can be explained by the sustained rise of income inequality following the economic crisis in Europe (Nolan, 2017). Rising income inequality combined with a high anti-globalization sentiment explain the success of left-wing populist political parties there (Burgoon, 2013). Hypothesis 4 which stated: ‘the higher the income inequality, the more likely an individual is to vote for a left-wing populist party’ is accepted.
To conclude, income inequality does not explain the success of populist political parties in Europe. This mechanism works only for Scandinavia. Income inequality does explain the success of left-wing populist parties. Income inequality increases negative attitudes towards immigration. These negative views on immigration explain the increasing number of votes for populist political parties.
Bibliography
Aggeborn, L. (2017). Public Finance and Right-Wing Populism. IFN working paper series, 1182 (1).
Bakker, R., De Vries, C., Edwards, E., Hooghe, L., Jolly, S., Marks, G., Polk, J., Rovny, J., Steenbergen, M., Vachudova, M.A. (2012). Measuring party positions in Europe: The Chapel Hill expert survey trend file, 1999 – 2010. Party Politics, 21(1), 1-15.
https://doi.org/10.1177/1354068812462931
Boros, T., Freitas, M., Stetter, E., & Kadlot, T. (2016). The state of populism; the state of populism in the European Union. Brussels, Belgium: Foundation for European progressive studies.
Burgoon, B. (2013). Inequality and anti-globalization backlash by political parties. European Union Politics, 14(3), 408–435. https://doi.org/10.1177/1465116512467546
Cuperus, R. (2017). The Populist Revolt Against Globalisation. Clingendael spectator, 71(1), 100–120.
Gidron, N., & Bonikowski, B. (2013). Varieties of Populism: Literature Review and Research Agenda. Weatherhead Center for International Affairs, Harvard University, Working Paper Series, No. 4, (13), 1–38. https://doi.org/10.2139/ssrn.2459387
Gidron, N. & Hall, P. A. (2017). The politics of social status: economic and cultural roots of the populist right. British Journal of Sociology, 68, S57–S84.
https://doi.org/10.1111/1468-4446.12319
Green, G. (2017). Economic Inequality as A Cause for Escalation in Populist Votes. Retrieved from https://www.behorizon.org/economic-inequality-populist-votes/ Guiso, L., Herrera, H., Morelli, M., & Sonno, T. (2018). Populism: Demand and Supply.
Mimeo, 1–63. Retrieved from http://www.heliosherrera.com/populism.pdf
Han, K. J. (2016). Income inequality and voting for radical right-wing parties. Electoral Studies, 42, 54–64. https://doi.org/10.1016/j.electstud.2016.02.001
Harteveld, E. (2016). Winning the “losers” but losing the “winners”? The electoral consequences of the radical right moving to the economic left. Electoral Studies, 44, 225–234. https://doi.org/10.1016/j.electstud.2016.08.015
Hellwig, T. (2012). Constructing Accountability : Party Position Taking and Economic Voting. Comparative political studies, 45(1), 91-118.
https://doi.org/10.1177/001041401142251
Inglehart, R. F. & Norris, P. (2016). Trump, Brexit, and the Rise of Populism: Economic Have-Nots and Cultural Backlash. Faculty Research Working Paper Series. Retrieved from http://ssrn.com/abstract=2818659
Kriesi, H. (2014). The Populist Challenge. West European Politics, 37(2), 361–378.
https://doi.org/10.1080/01402382.2014.887879
Kriesi, H., Grande, E., Dolezal, E., Helbling, M., Höglinger, D., Hutter, S., & Wüest, B. (2012). Political Conflict in Western Europe. Cambridge University Press.
Lewis-Beck, M. S., & Nadeau, R. (2011). Economic voting theory: Testing new dimensions. Electoral Studies, 30(2), 288–294. https://doi.org/10.1016/j.electstud.2010.09.001 Linn, S., & Nagler, J. (2017). Economic Voting and Economic Inequality: US Presidential
Elections 1952 – 2012. American Politics Research, 1–32.
https://doi.org/10.1177/1532673X16685313
Lucassen, G., & Lubbers, M. (2012). Who Fears What? Explaining Far-Right-Wing Preference in Europe by Distinguishing Perceived Cultural and Economic Ethnic Threats. Comparative Political Studies, 45(5), 547–574.
https://doi.org/10.1177/0010414011427851
Mudde, C. (2016). Europe’s populist surge: A long time in the making. Foreign Affairs, 95(6), 25–30.
Oesch, D. (2008). Explaining workers’ support for right-wing populist parties in Western Europe: Evidence from Austria, Belgium, France, Norway, and Switzerland.
International Political Science Review, 29(3), 349–373.
https://doi.org/10.1177/0192512107088390
Rooduijn, M. (2015). The rise of the populist radical right in Western Europe. European View, 14(1), 3–11. https://doi.org/10.1007/s12290-015-0347-5
Rooduijn, M., Burgoon, B., van Elsas, E. J., & van de Werfhorst, H. G. (2017). Radical distinction: Support for radical left and radical right parties in Europe. European Union Politics, 18(4), 536–559. https://doi.org/10.1177/1465116517718091
Rydgren, J. (2014). Radical Right-wing Populism In Sweden and Denmark. Ben Gurion University of the Negev, (March), 0–34. https://doi.org/10.1353/sais.0.0070
Winkler, H. (2014). The Effect of Income Inequality on Political Polarization: Evidence from European Regions, 2002 – 2014. SAIS review.
Appendices
Appendix I; list of ESS variables used
Variable(s) European Social
Survey question Variable European Social Survey question
Prtvtat, prtvtaat, prtvtbat
Austria, party voted
for in last election prtvtie, prtvtaie
Ireland, party voted for in last election Prtvtbe, prtvtabe,
prtvtbbe, prtvtcbe Belgium, party voted for in last election prtvtit, prtvtait, prtvtbit Italy, party voted for in last election Prtvtbg, prtvtabg,
prtvtbbg, prtvtcbg
Bulgaria, party voted for in last election
prtvtnl, prtvtanl, prtvtbnl, prtvtcnl, prtvtdnl, prtvtenl, prtvtfnl
Netherlands, party voted for in last election Prtvtcy, prtvtacy Cyprus, party voted
for in last election
Prtvtpl, prtvtapl, prtvtbpl, prtvtcpl
Poland, party voted for in last election prtvtcz, prtvtacz,
prtvtbcz, prtvtccz, prtvtdcz
Czech Republic, party voted for in last election
prtvtpt, prtvtapt, prtvtbp
Portugal, party voted for in last election Prtvtdk, prtvtadk,
prtvtbdk, prtvtcdk
Denmark, party voted for in last election
prtvtse, prtvtase, prtvtbse
Sweden, party voted for in last election Prtvtee, prtvtaee,
prtvtbee, prtvtcee, prtvtdee, prtvteee
Estonia, party voted
for in last election Prtvtasi, prtvtbsi, prtvtcsi, prtvtdsi, prtvtesi
Slovenia, party voted for in last election prtvtfi, prtvtafi,
prtvtbfi, prtvtcfi
Finland, party voted for in last election
prtvtes, prtvtaes, prtvtbes, prtvtces
Spain, party voted for in last election prtvtfr, prtvtafr,
prtvtbfr, prtvtcfr
France, party voted
for in last election Gincdif
Government should reduce differences in income levels Prtvde1, prtvde2, prtvade1, prtvade2, prtvbde1, prtvbde2, prtvcde1, prtvcde2, prtvdde1, prtvdde2, prtvede1, prtvede2
Germany, party voted
for in last election Imueclt
Country’s cultural life undermined or enriched by immigrants prtvtgb, prtvtagb,
prtvtbgb
Great Britain, party voted for in last election
imbgeco Immigration bad or good for country’s economy
prtvtgr, prtvtagr, prtvtbgr, prtvtcgr
Greece, party voted for in last election
imwbcnt Immigrants make
country worse or better place to live prtvthu, prtvtahu,
prtvtbhu, prtvtchu, prtvtdhu, prtvtehu
Hungary, party voted for in last election
Appendix II; list of populist political parties
Country Party name Left/Right
Austria Freedom Party of Austria (Fpo) Right
Belgium Vlaams Belang (vb) Right
Bulgaria Attack (Ataka) Right
Bulgaria Bulgarian National Movement (VMRO-BND) Right
Bulgaria Reload Bulgaria (BBT) Right
Bulgaria National Front for the Salvation of Bulgaria (NFSB) Right Bulgaria Citizens for European Development of Bulgaria
(GERB) Right
Cyprus Citizens’ alliance (SYM) Left
Czech Republic Dawn - national coalition (USVIT) Right
Germany National democratic party (NPD) Right
Germany Alternative for Germany (AfD) Right
Germany The Left (LINKE) Left
Denmark Danish People’s party Right
Estonia Conservative People’s Party of Estonia (EKRE) Right
Finland True Finns Right
France National Front (FN) Right
France Popular republican movement (MPF) Right
United Kingdom UK Independence Party (UKIP) Right
United Kingdom National Front (NF) Right
United Kingdom British National Party (BNP) Right
Greece Golden Dawn (XA) Right
Greece Independent Greeks (ANEL) Right
Greece Popular Orthodox Rally (LAOS) Right
Greece New Democracy (ND) Right
Greece SYRIZA Left
Hungary Jobbik movement Right
Hungary Fidesz Hungarian Civic Union Right
Italy Brothers of Italy (FdL) Right
Italy Northern league (LN) Right
Italy Five Star Movement (M5S) Left
Lithuania The Way of Courage (DK) Right
Luxembourg Alternative Democratic Reform (ADR) Right
Netherlands Party for Freedom (PVV) Right
Netherlands Reformed Political Party (SGP) Right
Netherlands List Pim Fortuyn (LPF) Right
Poland Law and Justice (PiS) Right
Poland United Poland (SP) Right
Poland Congress of the new right (KNP) Right
Slovakia Slovak National Party (SNS) Right
Slovakia Christian Democratic Movement (KDH) Right
Slovenia Slovenian Democratic Party (SDS) Right
Slovenia New Slovenia - Christian people’s party (NSI) Right
Spain PODEMOS Left
Appendix III; list of regions used
Region number
Region indication Countries
1 Northern, Western & Southern Europe Austria, Belgium, Cyprus, Germany, Denmark, Estonia, Spain, Great Britain, Finland, France, Ireland, Greece, Italy, Lithuania, Luxembourg, the Netherlands, Portugal, Sweden, Slovenia
2 Eastern Europe Bulgaria, Czech Republic, Hungary,
Poland, Slovakia
3 Northern & Western Europe Austria, Belgium, Germany,
Denmark, Estonia, France, Finland, Great Britain, Ireland, Lithuania, Luxembourg, the Netherlands, Sweden
4 Northern Europe Denmark, Finland, Ireland, Sweden,
United Kingdom
5 Scandinavia Denmark, Sweden
Appendix IV; country fixed included regression tables
Variables Populist vote
Gini 0.0615 (0.579) group(cntry) = 2, Belgium -0.990*** (0.00420) group(cntry) = 3, Bulgaria 0.388 (0.583) group(cntry) = 4, omitted -
group(cntry) = 5, Czech Republic -2.453**
(0.0136) group(cntry) = 6, Germany -3.111*** (0) group(cntry) = 7, Denmark -0.0128 (0.972) group(cntry) = 8, Estonia -3.048*** (1.42e-07) group(cntry) = 9, omitted - group(cntry) = 10, Finland -0.234 (0.605) group(cntry) = 11, France -0.691* (0.0911)
group(cntry) = 12, Great Britain -2.278**
(0.0160) group(cntry) = 13, Greece 0.835** (0.0413) group(cntry) = 14, Hungary 2.072*** (0) group(cntry) = 15, omitted - group(cntry) = 16, Italy 0.222 (0.648) group(cntry) = 17, omitted - group(cntry) = 18, omitted - group(cntry) = 19, Netherlands -0.177 (0.632) group(cntry) = 20, Poland 0.909*** (0.000545) group(cntry) = 21, omitted - group(cntry) = 22, Sweden -1.229** (0.0150) group(cntry) = 23, Slovenia 0.705
(0.185) group(cntry) = 24, Slovakia 0.0888 (0.915) Constant -3.893 (0.249) Observations 82,541
Robust pval in parentheses *** p<0.01, ** p<0.05, * p<0.1 (1) Variables immigration Gini -0.114** (0.0490) group(cntry) = 2, Belgium 1.157*** (0) group(cntry) = 3, Bulgaria 0.942*** (0.000703) group(cntry) = 4, Cyprus 0.169 (0.581)
group(cntry) = 5, Czech Republic -0.285
(0.270) group(cntry) = 6, Germany 0.866*** (0) group(cntry) = 7, Denmark 1.477*** (0) group(cntry) = 8, Estonia 1.420*** (0) group(cntry) = 9, Spain 1.601*** (7.64e-08) group(cntry) = 10, Finland 1.967*** (0) group(cntry) = 11, France 0.919*** (5.00e-07)
group(cntry) = 12, Great Britain 0.766***
(0.000326) group(cntry) = 13, Greece -0.406** (0.0391) group(cntry) = 14, Hungary 0.368*** (0.000114) group(cntry) = 15, Ireland 0.978*** (2.98e-10) group(cntry) = 16, Italy 0.483* (0.0570) group(cntry) = 17, Lithuania 1.814***
(3.81e-06) group(cntry) = 19, Netherlands 2.026*** (0) group(cntry) = 20, Poland 1.976*** (0) group(cntry) = 21, Portugal 1.325*** (0.00101) group(cntry) = 22, Sweden 2.212*** (0) group(cntry) = 23, Slovenia 0.177 (0.511) group(cntry) = 24, Slovakia -0.0966 (0.787) Constant 6.354*** (0.000304) Observations 146,921 Robust pval in parentheses
*** p<0.01, ** p<0.05, * p<0.1 (1) VARIABLES Popvoted immigration -0.165*** (0) group(cntry) = 2, Belgium 0.280 (0.461) group(cntry) = 3, Bulgaria 1.426* (0.0656) group(cntry) = 4, omitted -
group(cntry) = 5, Czech Republic -2.451**
(0.0105) group(cntry) = 6, Germany -1.498** (0.0160) group(cntry) = 7, Denmark 0.536** (0.0405) group(cntry) = 8, Estonia -2.474*** (0.000167) group(cntry) = 9, omitted - group(cntry) = 10, Finland 0.0398 (0.931) group(cntry) = 11, France 0.175 (0.566)
(0.0399) group(cntry) = 13, Greece 1.637*** (2.73e-08) group(cntry) = 14, Hungary 2.064*** (1.34e-05) group(cntry) = 15, omitted - group(cntry) = 16, Italy 0.0507 (0.942) group(cntry) = 17, omitted - group(cntry) = 18, Luxembourg -0.412 (0.135) group(cntry) = 19, Netherlands 0.525 (0.105) group(cntry) = 20, Poland 1.708*** (5.29e-06) group(cntry) = 21, omitted - group(cntry) = 22, Sweden -1.059* (0.0534) group(cntry) = 23, Slovenia 1.430*** (1.50e-05) group(cntry) = 24, Slovakia 0.557 (0.178) Constant -2.025*** (0) Observations 153,488
Robust pval in parentheses *** p<0.01, ** p<0.05, * p<0.1
