Master Thesis
The economic determinants of right-wing
populism in Europe
Abstract
After a long period of increasing globalization, the attitude towards globalization seems to be shifting in several parts of the world. Increasing populist sentiment and nationalism are creating pushback against global trade, and multinational governments. This thesis will focus on the rise of right-wing populism in Europe and will seek to analyse which factors drive people towards voting for right-wing populist parties. Previous literature on this topic has explored the institutional conditions for the formation of populist parties (the supply side) and only recently the attention has shifted towards the demand side. Using data from the European Social Survey this paper will further analyse the demand side of populism. The main findings are that anti-immigration attitudes, and Eurosceptic attitudes are positively related to right-wing populist voting. In addition, vulnerability to globalization is significant in Western-Europe. Unemployment, and financial distress are found to be insignificant determinants of right-wing populist voting.
Niels Frans S2543273 18 June, 2019
2
1. Introduction
Throughout history there have been two major globalization booms and one bust. The first period of globalization ended with the first world war, and the second period started after the second world war (Williamson, 2002). The years in between were a period of anti-globalization backlash. In recent years anti-anti-globalization sentiments seem to be on the rise again. Recent events such as the election of Donal Trump and the Brexit referendum have exemplified this. Trump’s election campaign was run on the famous “America first” slogan, and the United Kingdom seeks to leave the European Union to gain more autonomy. Both events indicate a clear shift towards a national focus and away form a global one. Both events have also been linked to populism, in the UK the United Kingdom Independence Party (UKIP), which can be classified as a right-wing populist party, played a significant in initiating the Brexit referendum. In other European countries right-wing populist parties are also on the rise with a corresponding increase of scepticism towards globalization (Rodrik, 2018). Figure 1 Gives an overview of the development of the voter share of populist parties in Europe between 1961 and 2015. In this graph a considerable rise in right-wing populism in Europe can be observed while left-wing populism has remained stagnant. The exact reason for this rise in right-wing populism in Europe remains a hotly debated topic.
3 This rise in right-wing populism begs the question as to why this phenomenon is occurring in multiple countries in Europe at the same time. Previous studies have given an overview populism from a global perspective (e.g. Rodrik, 2018) or analysed a specific country (e.g. Calontone and Stanig, 2016). However, a country perspective is not sufficient in explaining why right-wing populism is increasing throughout Europe. In his paper titled populism and
the economics of globalization Rodrik (2018) gives an overview of economic theories which
could explain the current rise of populism. One such theory is the Stolper-Samuelson theorem, which predicts that in addition to gains from trade, there will necessarily also be losers. He goes as far as to state that the anti-global backlash could have been predicted using economic theory. This paper will research whether the rise of right-wing populism in Europe could have been predicted by economic factors. Specifically, this will be done by analysing the demand side of populism by using individual level survey data from the European Social Survey (ESS). The demand side of populism focusses on grievances felt by the population which generate support for populism, while the supply side focusses on how populist parties emerge and position themselves (Mudde, 2007).
This paper will focus on the demand side for several reasons. Firstly, this is because the focus will be on the economic determinants of right-wing populist voting, and to answer this question it most suitable to focus on what drives individual voters instead of party
dynamics. To analyse this a macro-economic perspective or a micro-economic perspective can be taken. A macro-economic analysis would focus on macro-economic trends and how they impact right-wing populist voting, as is done in Lubbers et al. (2002) and Swank and Betz (2003). However, perhaps a more interesting route is to analyse it form a micro-perspective by focussing on an individual’s economic insecurity and attitudes. Secondly, previous studies on populism have extensively focused on institutional conditions which allow for the emergence of populist parties (the supply side) (Norris, 2005; Golder, 2016). Only recently have academics shifted their focus to the demand side of populism (Inglehart and Norris, 2016; Guiso et al., 2017; Boeri et al., 2018). It should be noted that many recent papers on the demand side of populism are currently unpublished and therefore care should be taken in interpreting results from these papers.
The study by Guiso et al. (2017) focusses on both left and right-wing populism as well as both the demand and the supply side. They find that their measure of economic insecurity is significant and that people who experience higher economic insecurity are more likely to vote for either left or right-wing populist parties. Inglehart and Norris (2016) focused on individual characteristics and cultural values and found that age, gender, education and authoritarian leanings all contribute towards the likelihood of voting for right-wing populist parties. A working paper by the IMF (Boeri et al., 2018) also uses data from the European Social Survey and focusses on the role of civil society as a predictor of right-wing populist voting behaviour. They find that gender, education, and most prominently, engagement in civil society have a significant effect.
4 to this I will also look at the role that globalization plays in driving support for right-wing populist parties. This will be done by analysing immigration, euroscepticism, and
vulnerability towards globalization.
In short this leads to the following research question: What are the (economic) determinants
of right-wing populism in Europe? This research question will be broken up into several
hypotheses which can be placed into two categories. The first set of hypotheses test whether financial distress, unemployment and vulnerability to globalisation increase the likelihood of voting for right-wing populist parties. The second set of hypotheses test whether dissatisfaction with country politics, dissatisfaction with the European Union, and dissatisfaction with immigration drive people towards voting for right-wing populist parties. In answering this question, I will first analyse the characteristics of right-wing populism and provide a clear definition. Then I will examine economic theory and empirics from previous papers in order to construct hypotheses to explain the rise of right-wing populism in Europe. The primary data source for this thesis will be the European Social Survey. Using this
dataset, I will cover 21 European countries, containing a total of 35,698 observations. The survey includes socio-demographics, political data, and economic data. In addition to the ESS data a party classification system from van Kessel (2015) will be used to determine which parties are classified as right-wing populist. A series of regressions will then be run to determine which factors increase the probability of an individual voting for right-wing populist parties.
The results in this paper further confirm several findings found in previous papers, but also contradict some. Contrary to Norris and Inglehart (2016), age and gender are not found to be significant factors in predicting voting behaviour. In addition, unemployment and financial distress are found to be insignificant predictors of right-wing populist voting in Europe. Perhaps the most interesting finding is that the measure of exposure to
globalization used in this paper is significant in Western-European countries, which is accordance with economic theory. This finding will be further elaborated upon in later sections. Furthermore, a person’s attitude towards immigration, albeit economically or culturally motivated, is a significant predictor of right-wing populist voting. Lastly, anti-EU sentiments is also a significant predictor of right-wing populist voting.
5
2. Literature review: Right-wing populism
2.1 Right-wing populism
In order to determine which factors play a role in right-wing populist voting behaviour it is important to first clearly define what populism and right-wing populism are. This is
particularly important because the definitions of populism and right-wing populism can vary depending of the angle of analysis. The following section will give a brief overview of
prominent literature on the topic of populism. The goal of this section is to highlight three characteristics of right-wing populism found in literature which are of interest for the hypotheses of this thesis. First; the anti-establishment nature of right-wing populism; Second, playing into the fears and grievances of the population; and third, the focus on nativism and homogeneity of the people. Why these characteristics are of importance will be highlighted in the following paragraphs.
The term populism first emerged in the United States when a coalition of miners and farmers fought against the Gold Standard (Rodrik, 2018). Since then the term has evolved and today it is used to describe a variety of political movements. This first instance of populism highlights the anti-establishment nature of populism. Guiso et al. (2017) use the encyclopaedia definition of populism as their starting point. This definition states that: 1) populist claim to promote the interest of the common citizens against the elites, and 2) pander to sentiments and fears of the people (Guiso et al., 2017). This component of populism can also be found in Acemoglu et al. (2013) who notes that populist often use rhetoric which aggressively defends the interest of the common man against the elites. Mudde (2007) describes populism as an opportunistic form of politics with the aim of pleasing the people, and thereby gaining the support of people for the next election cycle. These definitions illustrate that populist parties play into grievances felt by the population and pit them against the ‘elites’ of the established parties. Note that these two
characteristics apply to both left-and right-wing variants of populist parties.
What distinguishes right-wing populism from left-wing populism is the third characteristics, which is the focus on nativism and the homogeneity of the people (Mudde 2007; Inglehart amd Norris, 2016). The distinction between right-wing and left-wing populism can be explained as follows: right-wing populism is focused on cultural identity, while left-wing populism focusses on economic inequality and targets the rich and large multinational corporations (Guiso et al., 2017; Rodrik, 2018). The focus of right-wing populism on nativism is also pointed out by Mudde (2007), who emphasizes that right-wing populist often create distinction between the native population and others. It is for this reason that right-wing populism often criticizes immigrants and foreigners in their rhetoric. Using these
6 populist parties should make these party more prevalent for people who harbour anti-establishment and anti-EU sentiments. These characteristics of right-wing populism also mean that anti-immigration sentiments should play a large role in attracting people towards voting for right-wing populist parties. This is because on the one hand it plays into the economic grievances that people perceive as a result of immigration, such as labour market competition (Guiso et al., 2017), and welfare state erosion (Hatton, 2016; Schmidt and Spies, 2016). On the other hand, on the nativism side of right-wing populism this means that people who are opposed to immigration for cultural reasons are also more attracted
towards voting for a right-wing populist party.
These are not the only characteristics of right-wing populism. Other characteristics of populism which are mentioned in the literature, but which are not of central importance for this paper, are that leaders of right-wing populist parties have authoritarian leanings
(Inglehart and Norris, 2016), and that populist parties tend to disregard long-term consequences (Mudde; 2007; Guiso et al., 2016)
The three characteristics of most interest to this thesis are also embodied in the
classification of van Kessel (2015). Van Kessel (2015) has studied all political parties that have gained parliamentary representation in Europe between 2000 and 2013 and has identified the ones which can be characterised as populist. Van Kessel defines a party as populist if: 1) they portray ‘the people’ as virtuous and homogeneous; 2) advocate for popular sovereignty instead of elitist rule; 3) the party defines itself as being
anti-establishment (van Kessel, 2015). Van Kessel uses primary sources, such as party speeches and manifestos as a basis for his classification. In addition to this he uses a pool of experts to validate his classification. Using this method, he has identified 57 populist parties in 26 of the 31 examines European countries (Van Kessel, 2015). The fact that this classification is based on political strategy from primary sources, rather than subjective judgements makes this classification particularly useful. Because of this, Van Kessel’s (2015) classification will be used in this paper to identify populist parties.
2.2 Economic determinants of right-wing populism
Having covered the literature on populism, I will now move to the theoretical and empirical literature on the economic determinants of right-wing populism. On the demand side of populism, a recent paper by Inglehart and Norris (2016) examined the rising support for right-wing populist parties in Europe by analysing individual characteristics. This paper used the European Social Survey to test two theories by which the rise of right-wing populism can be explained. Firstly, through the economic security/inequality theory. In this theory
7 which contains 293,856 respondents covering 32 countries (Inglehart and Norris, 2016). A multivariate logistic regression is used to find evidence for their theories. They found little support for the economic insecurity theory and found the cultural backlash theory to be most prevalent. The cultural backlash theory suggests that the increase in votes for right-wing populist can largely be explained as backlash against rapid progressive cultural change. The full extent of the cultural backlash thesis is beyond the scope of this paper, but I will include cultural variables relating to anti-establishment and anti-global values.
In contrast to Inglehart and Norris (2016), Algan et al. (2018) found that economic insecurity does play a substantial role in explaining the rise of right-wing populism in Europe. Algan et al. (2018) used regional unemployment data from Eurostat, covering 217 regions in 25 countries. They combined this with parliamentary election data and data from the European Social Survey. By doing this they found that economic insecurity and unemployment are significant predictors of right-wing populist voting when accounting for time-invariant factors (Algan et al. 2018). In addition, they argue that cultural attitudes, such as
institutional distrust, are a result of economic insecurity and not independent drivers. Other literature also found empirical evidence for the economic argument on the rise of right-wing populism. Lubbers et al. (2002) studied the voting behaviour of extreme right-wing parties in Europe. Note that extreme-right wing is not synonymous with right-wing populism, but nevertheless their study can provide some insight into potential economic factors at play. They made use of large-scale survey data containing 49,801 observations and country-level macro statistics. They report that support for extreme-right wing parties is more prevalent among unemployed people and among uneducated people.
There are also academics who doubt the importance of economic arguments behind the rise of right-wing populism. Kitschelt (2002) argues that the socio-economic profile of right-wing populist voters is much more complex than the stereotype of low-skilled, economically insecure, and unemployed people, and that economic explanations are insufficient. Mudde (2007) similarly opposes the idea that economic explanations are sufficient in explaining the rise of right-wing populism.
8 2.2.1 Unemployment
There are multiple papers which suggest a link between unemployment and support for right-wing populism. Studying the political consequences of the Financial Crisis in Europe, Algan et al. (2018) found that in elections after 2008 a decline in trust in established politics and political institutions was sharpest in regions which were most affected by
unemployment. This distrust in political institutions in turn causes people to vote for right-wing populist parties (Algan et al., 2018). This is also supported by Guiso et al. (2017) who found that people who have low trust in political institutions are more inclined to vote for right-wing populist parties or to abstain from voting. Like this paper, Guiso et al. (2017), use the European Social Survey as their source of data. They used seven ESS waves from 2002 to 2014, containing a total of 134,834 observations across 24 European countries. From this dataset they constructed a pseudo-panel to measure the development of right-wing
9 To summarize, there is no unanimous conclusion in the literature on whether
unemployment influences voting behaviour, and in which direction this effect goes. In literature where a positive correlation between unemployment and right-wing populism was found the channel through which unemployment affects right-wing populism is as follows: unemployment causes a loss of trust in the political establishment within the country which then in turn causes people to vote for right-wing populist parties. This leads to my first hypothesis:
Hypothesis 1: Unemployment increases the probability of voting for a right-wing populist party.
2.2.2 Financial distress
In addition to unemployment, another key component of the economic insecurity theory brought forth by Inglehart and Norris (2016) is that rising financial distress causes an
increase of support for populist parties. The theory behind this is that financial distress fuels distrust and resentment towards the ruling political parties and makes people more
susceptible to right-wing populist parties offering an alternative (Inglehart and Norris, 2016). These results are consistent with Guiso et al. (2017) who also predict that economic insecurity influences right-wing populist voting. On the demand side of populism, they found that the people who face the greatest income difficulties are the people who are most prone to abstain for voting or vote for a populist party. In addition, Guiso et al. (2017) argue that when shocks to economic security occur a sharp negative change in political trust towards the government and towards immigration occurs. In addition to this Guiso et al. (2017) also found that economic insecurity negatively affects a person’s trust in the political system, and thereby causes people to abstain from voting. This abstinence from voting indirectly benefits right-wing populist parties, because votes are taken away from established political parties and thereby proportionally increase the voter share of right-wing populist parties (Guiso et al., 2017). This was also found by Anayev and Guriev (2016) on a study of the 2009 recession in Russia. They found that a decrease in GDP causes a decline in trust in people, political parties, institutions and politicians, because people blame the government for poor economic performance.
In addition to financial distress influencing people’s distrust in their national government, it could also negatively impact their attitude towards European Institutions. A research by Dustmann et al. (2017), using survey data from the ESS and regional parliament election data, find that in the aftermath of the European Financial Crisis, which resulted in a significant rise in economic insecurity, distrust of European institutions increased. This distrust could largely be explained by poorer economic and was found to be correlated with an increase in populist voting (Dustmann et al., 2017).
10 to abstain from voting or to vote for right-wing populist parties. This brings me to the
following hypothesis:
Hypothesis 2: Higher personal financial distress increases the probability of voting for a right-wing populist party.
2.2.3 Vulnerability to globalization
The economic insecurity perspective has also been discussed in a recent paper by (Rodrik, 2017), in which he gives an overview of various historic and economic reasons which contribute to a rise in populism. Rodrik (2018) refers to the Stolper-Samuelson theory on trade to illustrate how globalization through free-trade can also create losers. The Stolper-Samuelson theorem is a model with two goods and two factors of production, with full labour mobility of the factors. The factor which is used intensively in the importable goods must experience a decline in real earnings (Rodrik, 2018). In the case of Europe, the Stolper– Samuelson models predicts that unskilled workers will see a loss in employment and
earnings as a result of globalization. This is because the theory suggest that the relative prices commanded by comparatively scarce factors in the developed economies (unskilled workers) decline with internationalization as the relative demand for comparatively abundant factors (high-skill workers) increases (Swank and Betz, 2003).
The argument that globalization causes some people to be worse off is illustrated by Gereffi (2014). Because of the prevalence of Global Value Chains manufacturing jobs are
transferred from the developed economies to emerging economies (Gereffi, 2014). Autor et al. (2018) in their paper on the China trade shock, analyses how the China trade shock created losers from trade, particularly in specific geographical areas of the United States. The China trade shock also caused labor-market disruption and lead to a loss of
manufacturing jobs (Autor, 2018). Timmer et al. (2012) find in their research of 27 European countries between 1995 and 2008 that through participation in the Global Value Chain there is a large job increase in services, but a job loss in manufacturing. Timmer et al. (2012) also note a shift away from low-skilled manufacturing workers towards high-skill workers. Interestingly, opposition to trade is much less prevalent in Europe than the United States. A potential reason for this could be that welfare states are generally much more developed in Europe, which negates the negative effects of international trade through redistribution. Consequently, the gains of trade are better distributed in Europe, which causes less opposition towards international trade (Rodrik, 2018). This explanation by Rodrik can be empirically supported by a study by Swank and Betz (2003), who found that a more developed welfare state is more able to redistribute the benefits from trade among the population. Therefore, direct opposition towards trade will find little support for right-wing populist parties in Europe.1
11 Literature on globalization find that the people who are most vulnerable to globalization are people who work as low-skill workers in low-tech manufacturing industries. As a result, these people could feel let down by the established political parties and will turn to right-wing populist parties. Based on this I created the folloright-wing hypothesis:
Hypothesis 3: People who are more exposed to globalization are more likely to vote for right-wing populist parties.
2.3 Cultural determinants of right-wing populism
Many academics argue that the channel through which economic insecurity affects right-wing populist voting, is due to a loss of trust in established political. This interconnection between economic determinants and cultural attitudes make it interesting to test both theories. The immigration can be considered both an economic factor and a cultural factor. In this case immigration is part of the cultural theory, although it could also have been included in the economic insecurity theory. In this section I will examine literature on the effect of the following factors on right-wing populist voting: immigration, anti-establishment sentiments, and anti-global sentiments.
2.3.1 Immigration
Another potential predictor of right-wing populist voting behaviour is a person’s attitude towards immigration. A negative attitude towards immigration could be driven by either cultural values or through economic insecurity (Guiso et al., 2017). In the case of cultural values this directly influences the likelihood of voting for right-wing populist parties, as one of the characteristics of right-wing populist parties is a focus on nativism. Economic
insecurity as a result of immigration indirectly affects right-wing populist voting. In this case people perceive their economic insecurity as a consequence of immigration and therefore are more likely to vote for right-wing populist parties. (Guiso et al., 2017)
12 Immigration does not only cause cultural backlash but can also have negative economic effects. In a study on the effect of immigration on the welfare state in Germany, Schmidt and Spies (2016) found that the native population becomes less supportive of welfare programs as the proportion of immigrants increases at the regional level. Moreover, they found that this effect is further strengthened by a negative economic context2. These results
are supported by Hatton (2016), who found that a large concern for right-wing parties in Europe is that a high amount of immigration erodes welfare state benefits (Hatton, 2016). right-wing populist parties also find more support from voters when they perceive increased competition from immigrants in public housing (Cavaille and Ferwerda, 2017).
2.3.2 Anti-establishment and Euroscepticism
In section 2.1 it was covered how anti-establishment sentiments are a key characteristic of right-wing populist parties. In addition, numerous studies suggest a relationship between right-wing populism and anti-establishment rhetoric. Findings by Guiso et al. (2017) suggest that people who have the lowest trust in traditional political parties and governmental institutions are also most likely to be convinced by anti-elite populist rhetoric. There seems to be a relation between unemployment, distrust in the political establishment, and support for right-wing populism. A study by Sitter (2001) suggest that euro-scepticism Is another form of government-opposition dynamics (Sitter, 2001). These factors will be discussed together in this section.
A prominent example of right-wing populism against global institutions is Brexit. Norris and Inglehart (2016) as well as Becker et al. (2016) seek to explain the voter behaviour which led to Brexit. Unsurprisingly, they found that anti-EU sentiments, as well as populism played a role in the Brexit vote. This is because, the United Kingdom Independence Party which can be classified as a right-wing populist party, played a significant in instigating the Brexit referendum. In addition, Becker et al. (2016) found a strong link between anti-immigration sentiments and voting for Brexit. They also find that older people, people who are lower educated, and areas with low pay and high unemployment were more inclined to Vote Leave.
Interestingly, Rooduijn (2017) finds that euroscepticism and distrust in political parties do not seem to influence a person’s probability of voting for populist parties. However, it should be noted that Rooduijn focuses on populism in general instead of solely right-wing populism, which could be a potential reason why no relation was found. Based on these findings I have set up the final two hypotheses:
Hypothesis 5: People who have are more dissatisfied with political parties are more likely to vote for right-wing populist parties
13
Hypothesis 6: People who are eurosceptic are more likely to vote for right-wing populist parties.
3. Methodology
The following sections will contain the model and the empirical strategy of this paper. I will start by explaining the econometric specifications in this paper, containing the motivation of the 3 main specifications and basic information on the variables used. A more detailed explanations of the variables discussed in the models will be given in the data section. I will also cover several econometric problems of the model and how this could affect the results.
3.1 The model
The dependent variables in this model will be support for Right-Wing Populist Parties (RWPP), and this will be a binary variable. When dealing with a binary dependent variable there are three main methods; a linear probability model (LPM), a logit model, and a probit model. For the main specification in this paper I have chosen the probit model. Both logit and probit models are suitable for models estimating binary dependent variables, however they are more difficult to interpret than an LPM model. I have chosen the probit model as the main specification but are few substantial differences in the results between the three methods. When presenting the results from a probit model I will report the marginal effect of each variable. The marginal effect will report the estimated change in probability of voting for a right-wing populist party, while holding all other variables constant at the mean. By doing this the marginal effects reported will represent the estimated effect that each explanatory variable has on the probability that the average individual will vote for a right-wing populist party. Similar methods of reporting the marginal results of a probit model are used in Rodrik and Mayda (2005), and Boeri et al. (2018). The results of both the LPM and Logit models will be shown in the robustness check section.
Three specifications will be discussed in this section, the first specification will test the effect that the economic variables have on right-wing populist voting. The second specification will include the cultural variable of ant-establishment, anti-EU attitudes and immigration. Lastly, the third specification will include all variables, including country dummies and controls. The first model specification measures the economic determinants of right-wing populism and looks as follows:
1) 𝑅𝑊𝑃𝑃𝑖,𝑐 = 𝛼𝑖 + 𝛽1𝑢𝑛𝑒𝑚𝑝𝑖,𝑐+ 𝛽2𝑓𝑖𝑛𝑑𝑖𝑠𝑡𝑖,𝑐+ 𝛽3𝐿𝑆𝐿𝑇𝑀𝑖,𝑐+ 𝛽4𝑎𝑔𝑒𝑖,𝑐+
𝛽5𝑔𝑒𝑛𝑑𝑒𝑟𝑖,𝑐+ 𝛽6𝑒𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛𝑦𝑒𝑎𝑟𝑠𝑖,𝑐+ 𝛽7𝑚𝑖𝑛𝑜𝑟𝑖𝑡𝑦𝑖,𝑐+ 𝛽8𝑙𝑒𝑓𝑡𝑟𝑖𝑔ℎ𝑡𝑠𝑐𝑎𝑙𝑒𝑖,𝑐+ 𝜀𝑖,𝑐
14 interest in this specification. Firstly, unemp, which refers to whether a person has
experienced a period of unemployment of longer than 3 months in the last 5 years. Secondly, findist represents the subjective sense of financial distress that an individual experiences. And third, LSTM refers Low-skill Low-Tech Manufacturing, meaning that an individual is employed as a low-skill person in a sector classified as low-tech manufacturing. The remaining variables are control variables and include age, and gender, which takes the value of 0 if the respondent is male and 1 if the respondent is female. Furthermore, it includes the control variable educationyears which is a proxy for a person’s education level measured in years of completed education. Minority refers to an individual’s status as belonging to a minority ethnic group in a country. Lastly, leftrightscale controls for a person’s self-reported political orientation on a 1(left) to 10(right) scale.
2) 𝑅𝑊𝑃𝑃𝑖,𝑐 = 𝛼𝑖 + 𝛽1𝑡𝑟𝑠𝑡𝑝𝑎𝑟𝑡𝑖,𝑐+ 𝛽2𝑡𝑟𝑠𝑡𝑒𝑝𝑖,𝑐+ 𝛽3𝑖𝑚𝑚𝑖𝑔𝑟𝑖,𝑐+ 𝛽4𝑎𝑔𝑒𝑖,𝑐 + 𝛽5𝑔𝑒𝑛𝑑𝑒𝑟𝑖,𝑐+ 𝛽6𝑒𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛𝑦𝑒𝑎𝑟𝑠𝑖,𝑐+ 𝛽7𝑚𝑖𝑛𝑜𝑟𝑖𝑡𝑦𝑖,𝑐+ 𝛽8𝑙𝑒𝑓𝑡𝑟𝑖𝑔ℎ𝑡𝑠𝑐𝑎𝑙𝑒𝑖,𝑐+ 𝜀𝑖,𝑐
The second model specification features the anti-establishment and anti-globalization explanatory variables. The first explanatory variable trstpart is a measure of an individuals trust in the country’s political parties on a scale from 1 to 10. The second explanatory variable trstep is a measure of trust in the European Parliament and follows the same scale. The third variable immigr is a measure of attitude towards immigration from poorer
countries outside of Europe. This variable is included in the second specification because the labour market displacement effect and the cultural backlash effect of immigration are difficult to disentangle. The remaining variables control variables in the specification are the same as in the first specification.
3) 𝑅𝑊𝑃𝑃𝑖,𝑐 = 𝛼𝑖 + 𝛽1𝑢𝑛𝑒𝑚𝑝𝑖,𝑐+ 𝛽2𝑓𝑖𝑛𝑑𝑖𝑠𝑡𝑖,𝑐+ 𝛽3𝐿𝑆𝐿𝑇𝑀𝑖,𝑐+ 𝛽4𝑡𝑟𝑠𝑡𝑝𝑎𝑟𝑡𝑖,𝑐+
𝛽5𝑡𝑟𝑠𝑡𝑒𝑝𝑖,𝑐+ 𝛽6𝑖𝑚𝑚𝑖𝑔𝑟𝑖,𝑐+ 𝛽7𝑎𝑔𝑒𝑖,𝑐+ 𝛽9𝑔𝑒𝑛𝑑𝑒𝑟𝑖,𝑐+ 𝛽10𝑒𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛𝑦𝑒𝑎𝑟𝑠𝑖,𝑐+ 𝛽11𝑚𝑖𝑛𝑜𝑟𝑖𝑡𝑦𝑖,𝑐 + 𝛽12𝑙𝑒𝑓𝑡𝑟𝑖𝑔ℎ𝑡𝑠𝑐𝑎𝑙𝑒𝑖,𝑐+ 𝜀𝑖,𝑐
Lastly, I will include a specification which will include both the economic model and the anti-establishment/globalization model, this is shown in equation 3.
3.2 Data
3.2.1 Data and econometric problems
15 there is indeed a country effect which affects all individuals in a country in a similar manner, such as a recession, this effect will be captured by the country-dummy variable. A similar method of controlling for such effects is used in (Rodrik and Mayda, 2005; Inglehart and Norris, 2016; Guiso et al., 2018).
Another potential problem with applying a standard probit model is an endogenous selection issue. This issue arises because whether a person votes for a right-wing populist party is conditional on whether this person votes in the first place (Guiso et al., 2017). order to remedy this a methodology similar to Guiso et al. (2017) could be applied, where two-step Heckman probit model and an instrumental variable were used to account for this. From the empirical and theoretical literature, it seems that many of the explanatory variables are interconnected to a certain degree. Therefore, it is important to test for multicollinearity in order to ensure that the explanatory variables are not too highly correlated. A matrix of the correlations between explanatory variables can be found in appendix B and show that the correlations are not high enough to cause multicollinearity problems.
3.2.2 Explaining the variables
For the classification of right-wing populist parties I will use van Kessel’s (2015) party classification. Briefly summarized, van Kessel (2015) defines a party as populist when it meets three criteria. 1) If it portrays the people as virtuous and homogeneous; 2) if It advocates for popular sovereignty instead of elitist rule; and 3) if the party defines itself as outside of the political establishment. While the qualification of populist parties is always subjective to a certain degree, I chose van Kessel’s classification for several reasons. Firstly, van Kessel uses party manifestos and speeches as his primary source. And secondly, he uses local experts from each country to validate his classifications through answering a
questionnaire. By using the party manifestos and speeches as the primary source and then using experts to validate his classification he keeps subjective judgement to a minimum. The ESS contains questions on whether people voted in the last parliamentary election in their own country and which country they voted for. Using van Kessel’s qualification of populist parties, I am then able to create a dummy variable which takes the value of 1 if the person voted for a populist party, and 0 if not. An overview of the countries and parties included, and their classification can be found in table 1.
Because the voting data is derived from a survey instead of data instead of directly from the voting booth there could be some deviation between what people answered in the survey and their actual vote. This could be for several reasons, for example; people may be
reluctant to truthfully answer the survey questions, or people could just misremember. In a similar research by Guiso et al. (2017), which uses ESS rounds 1-7 and also used van Kessel’s classification of populist parties, they found that there tends to be a systematic
16 Table 1: Right-wing populist party classification
Country Party Last election Austria Freedom Party (FPÖ)
Alliance for the Future of Austria (BZÖ) Team Stronach (TS)
2013 2008 2013 Belgium Flemish Interest (VB)
National Front (FN) List Decker (LDD)
2010 2010 2010 Czech Republic ANO 2011 (ANO)
Dawn of Direct Democracy (Usvit)
2013 2013 Denmark Danish People’s Party (DF) 2011 Estonia -
Finland True Finns (PS) 2011 France National Front (FN) 2012 Germany -
Hungary FIDESZ- Hungarian Civic Alliance (MPSZ) Movement for a Better Hungary (Jobbik)
2010 2010
Ireland Sinn Féin 2011
Lithuania Order and Justice Party (TT) 2012 Netherlands Freedom Party (PVV) 2012 Norway Progress Party (FRP) 2013 Poland Law and Justice (PiS) 2011 Portugal -
Slovenia Slovenian National Party (SLS) 2011 Sweden Sweden Democrats (SD) 2010 Switserland Swiss People Party (SVP)
League of Tinicesians (LdTi)
2011 2011
Table 1: Right-wing populist parties in Europe who gained parliamentary representation after the latest national election. Source: based on van Kessel (2015), and ESS wave 7.
After having classified which parties are right-wing populist parties it is possible to view the share of votes that they have received. Figure 2 shows the percentage of votes received by right-wing populist parties per country. The most notable result is that in Hungary the populist share of votes account for more than 30 percent of the voter share. We can also see that Eastern-European countries such as Poland, the Czech Republic and Hungary have a larger share of RWPP votes than Western-European countries. It is also notable that in Germany, Estonia and Portugal no right-wing populist parties are present according to the van Kessel (2015) classification. This is especially notable for Germany, because the party Alternative for Deutschland (AfD) is often associated with populism. However, the AfD was founded in 2013 and did at the time not meet the criteria to be considered a right-wing populist party.
17 RWPP in the dataset. Secondly, a different classification for right-wing populist parties was used, which could be more lenient in identifying right-wing populist parties. And thirdly, the dataset contains different countries, which would lead to different measures of voter shares.
Figure 2: Share of RWPP votes per country: Source: ESS round 7
18 behaviour. In the robustness section the results using the first measurement of
unemployment will also be reported.
The second economic determinant of right-wing populist behaviour is a measure of financial distress. The ESS asks the question: Which of the descriptions comes closest to how you feel
about your household’s income nowadays. The responses to this question are measure on a
scale of 1 to 4 with the following values: 1) Living comfortably on present income, 2) coping on present income, 3) difficult on present income, 4) very difficult on present income. Many researchers have measured the decline of manufacturing in developed countries as a result of globalization (Gereffi, 2014; Autor, 2016), and often find that low-skill
manufacturing workers in low-tech industries are susceptible to the negative consequences of globalization (Timmer et al., 2012). To measure whether people who suffer the
consequences of globalization are more likely to vote for right-wing populist parties I created the Low-Skilled, low-Tech Manufacturing variable. This is done by taking the results from the question which measures the respondent’s education level on the ISCED scale3. If
the respondent has a education level below ISCED 2 they can be classified as low-skill UNESCO (2012). In addition, I created a dummy variable which takes the value of 1 if the respondent works in a low-tech manufacturing job, by using the NACER2 classification system4. If the respondent is both low-skilled and works in a low-tech manufacturing
industry (s)he will be classified as Low-Skill, Low-Tech Manufacturing (LSLTM). This will result in a dummy variable which takes the value of 1 if a respondent meets the LSLTM requirements. By doing this I can capture the effect of exposure to globalization to a certain degree
To further test this theory, I will also create a dummy variable for Western-European countries. Following the Stolper-Samuelson theorem comparatively scarce factors in the developed economies (unskilled workers) decline with internationalization as the relative demand for comparatively abundant factors (high-skill workers) increases (Swank and Betz, 2003). Thus, in more developed Western-European countries, low-skilled manufacturing workers should be more impacted by globalization than their counterparts in Eastern-European countries.
Lastly, I will measure whether immigration affects the probability of a person voting for a right-wing populist party. I will choose the variable which most closely represents intended meaning behind the explanatory variable. For example; there are multiple questions on immigration, some of which refer to the ethnicity of immigrants, and others to the economic consequences of immigration. Both of these questions in essence measure the respondent’s attitude towards immigration and could both be used to measure anti-immigration sentiments. This is reflected in the fact that choice of proxy for anti-immigration attitudes has no effect on the significance of the anti-immigration sentiment variable. The
3 The International Standard Classification of Education (ISCED) is a statistical framework for classifying education by UNESCO.
19 results from other measurements of immigration are shown in the robustness section. The measurement of immigration used in the main specifications is based on the following ESS question how people feel about immigrant from poorer countries outside Europe. The answers range from 1 (Allow many to come and live here) to 4(Allow none).
To isolate the effect of the explanatory variables on right-wing populist voting behaviour I will also include several relevant control variables. The decision to include these variables is based on previous literature. Firstly, I will include the standard control variables of age and gender. Multiple studies have shown that these factors are relevant to voting behaviour (Lubbers et al., 2002, Rooduijn, 2014). I will also control for education by using the proxy ‘years of completed education’ (Inglehart and Norris, 2016; Guiso et al. 2017). As discussed in the section on defining right-wing populism, belonging to the majority or native group could also play a role in the voting behaviour for right-wing populist parties. Therefore, I will use the ESS question: Do you belong to a minority ethnic group? to control for the ethnicity of the respondent. Lastly, I want to measure how likely people are to vote for a right-wing populist party conditional on their current political orientation. This is measured in a question of the ESS where people self-report their political orientation on a scale from 0(Left) to 10(Right).
3.2.3 Summary statistics
In order to get a feel for the data used in this paper I have provided a graph of summary statistics of all previously mentioned variables in table 2.
Table 2: Descriptive statistics
N Mean Std. Dev. Minimum Maximum
Dependent variable RWPP 35,698 0.0794162 0.2703911 0 1 Control variables Age in years 35,623 49.43882 18.67469 14 114 Gender (0=men, 1= women) 35,676 1.531366 0.4990222 1 2 minority 35,277 1.942399 0.2329912 1 2 Education years 35,418 12.90132 3.839232 0 50 Left-rightscale 31,794 5.069824 2.164346 0 10 Economic variables Unemployment 9,733 1.536833 0.4986671 1 2 Financial distress 35,370 1.929573 0.8204731 1 4 LSLTM 35,698 0.0580705 0.23388 0 1 Cultural variables
Trust in political parties 35,075 3.678204 2.381962 0 10 Trust in EU parliament 33,529 4.251931 2.507463 0 10 Immigration 34,745 2.59905 0.9143207 1 4
20 The dependent variable RWPP shows a mean of approximately 8 percent. This is notably lower than the 20 percent populism in Europe shown earlier in figure 1.
The variable unemployment has a mean of 1.53 which means that around 44 percent of people have experienced unemployment for a period of more than 3 months in the last 5 years. Unfortunately, the unemployment variable suffers from a large amount of missing observations. This brings the amount of observations in included in the regression down from approximately 34,000 to approximately 7,000.
For financial distress, approximately 20 percent of the respondents reported to be have some difficulty or a lot of difficulty with living on their present incomes. For LSLTM
approximately 5 percent of the respondents met the criteria of being a low-skilled worker in a low-tech manufacturing industry. Anti-establishment attitudes has a mean value of around 3.7 while answer range from 1 to 10. This indicates that a majority of the people report that they have below average trust in the political parties in their country.
Interestingly, the mean for trust in the European Parliament is higher at 4.3. This indicates that respondents on average have a slightly higher trust in the European Parliament than their national parties. However, the trust in the European Parliament is still generally below average. For immigration the mean is around 2.5, which means that most people are more negative towards immigration than positive. This can be seen with any measure of
21
4. Results
Table 3: Estimation results
Specification 1)Economic 2)Cultural 3)Total 4)West 5)East
Age 0.000158 0.0000214 0.0000321 -0.000138 0.000971** (0.000195) (0.000244) (0.000221) (0.000222) (0.000374) Gender -0.0129 -0.00955 -0.0104 -0.0141 0.00472 (0.0112) (0.0116) (0.0116) (0.0126) (0.0231) Minority -0.0579* -0.0501** -0.0515** -0.0966*** 0.0237 (0.023) (0.0194) (0.0197) (0.0225) (0.0247) Education -0.00394** -0.00430** -0.00364* -0.00297* -0.00403 (0.00152) (0.00152) (0.00143) (0.00145) (0.00441) Left-right 0.0231*** 0.0231*** 0.0233*** 0.0155*** 0.0417*** Scale (0.00461) (0.00443) (0.00429) (0.00374) (0.00856) Unemployment -0.000327 0.000593 -0.000732 -0.00127 (0.00546) (0.00574) (0.00547) (0.0138) Financial 0.0103 0.0102 0.00879 0.0029 distress (0.00546) (0.0054) (0.0059) (0.0117) LSLTM 0.0194 0.0233 0.0471*** -0.0186 (0.015) (0.0157) (0.0143) (0.0134) Trust in 0.00226 0.00252 -0.00171 0.0150** Political parties (0.00307) (0.00302) (0.00244) (0.0052) Trust in EP -0.0111*** -0.0110*** -0.0105** -0.0102* (0.00276) (0.00274) (0.00349) (0.00441) Immigration 0.0381*** (0.00863) 0.0348*** (0.00795) 0.0343*** (0.00800) 0.0361*** (0.00933) 0.0300* (0.0141) Pseudo R2 14.43% 17.09% 17.25% 14.83% 20.01% Observations 6880 6768 6453 4744 1709
Notes: Table contains the estimated marginal effect of the probability of voting for a RWPP, given an increase in the value
of the explanatory variable, while holding all other explanatory variables at their mean value. The standard error of the marginal effects (adjusted for clustering on country) are reported below the marginal effects. The significance level is represented by *** for 0.001, ** for 0.01, * for 0.05.
Table 3 shows the following specifications: (1) economic model (2)
Western-22 European countries (5) the total model with only Eastern-European countries5. All models
include a full set of control variables and country dummies. For clarity, the country dummies are not shown in table 3, but in appendix A.
Starting with the socio-demographic controls we can see that contrary to findings in previous papers (Inglehart and Norris, 2016; Guiso et al.,2017) age does not seem to be a significant predictor of support for voting for right-wing populist parties. Only in model (5), which is the Eastern-European model we can see that age is a significant factor. Similarly, gender also does not seem to play a significant role in predicting an individual’s likelihood of voting for a right-wing populist party. On the other hand, belonging to a minority group has a significant effect on predicting voting behaviour. In all cases where the status as a minority is significant the coefficient is negative. This is in line with previously discussed description of right-wing populist parties, where nativism and belonging to a majority group were found to be recurring characteristics in classifying right-wing populist parties. The result also show education to be a significant factor in predicting right-wing populist voting. In all cases the coefficient is negative, which indicates that right-wing populist parties receive more support from people who are less educated. Unsurprisingly, belonging to an ethnic minority is also negatively correlated with RWPP. This is unsurprising, given the focus on nativism and the majority group of right-wing populist parties. This also confirms the findings of previous papers that education and minority status are significant factors in RWPP voting. The exact underlying reasons for this are up for discussion and could be a focus of further research. It is also noteworthy that in contrast to the other specifications, the Eastern European
specification does not find education and minority status to be significant factors.
Unfortunately, this paper also does not provide any explanation for the differences between Western and Eastern Europe. The most likely explanation is that the relatively large
economic and cultural differences between Western-and Eastern European countries affects the way in which people react to right-wing populist parties. However, this explanation is just speculation and further research should be done into this topic to provide substantiated explanation.
Now moving to the real hypothesis of the paper; whether economic determinants can predict right-wing populist voting behavior. Starting with unemployment the results in all specifications are insignificant. Unfortunately, no support for the hypothesis that
experiencing unemployment increases the probability of voting for a right-wing populist party. However, this result is not entirely unsurprising, since previous studies have found contradictory results which indicate that unemployment both increases and decreases support for right-wing populist parties. The studies which found unemployment to be positively correlated with RWPP argue that being unemployed causes loss of trust and resentment towards mainstream parties increases the likelihood of voting for RWPP (Guiso
23 et al., 2017, Boeri et al., 2018) . On the other hand (Knigge, 1998) argues that during times of serious economic problems people will look to established parties to solve the problem. The second economic variable, financial distress also appears to be insignificant across all variations. Consequently, no support for the hypothesis that economic insecurity is a predictor of right-wing populist voting can be found. This result is surprising, considering that papers which use a similar subjective measure of financial distress (Norris and
Inglehart, 2016; Guiso et al., 2017) report that financial distress is a significant measure of RWPP. The variable is insignificant in all main specifications and is also found to be
insignificant in the multiple robustness checks. With both unemployment and economic security being insignificant these results aligns most closely with the conclusion drawn by Rooduijn (2017), who found that the ‘populist voter’ does not exist.
Perhaps the most interesting explanatory variable is LSLTM, which captures the vulnerability to globalization. In model 1, 2, 3 and 5 this variable is insignificant, and it is only significant in specification 4. While this does not confirm the hypothesis that people in skill low-tech manufacturing jobs in Europe are more likely to vote for right-wing populist parties, it does not mean that no conclusions can be drawn from this variable. The fact that the LSLTM variable is only significant in model 4, which only includes Western-European countries is entirely in line with the predictions of the Stolper-Samuelson theorem.
Trust in political parties is also mostly insignificant. This result is surprising, as it contradicts one of the core characteristics of right-wing populist parties, and it contrasts with findings form other papers. On the other hand, euroskeptisism seems to be a highly significant predictor of RWPP voting in specification 2,3, and 4. Lastly, the immigration variable is highly significant across all specifications. This is likely due to the fact that immigration includes both economic and cultural arguments for voting for RWPP.
4.1 Robustness check
To test the robustness of the model I will run the total model with several modifications. The results of these modifications can be seen in table 4. Specification (1) and (2) in table 4 show the results of the Logit model and LPM model respectively. For the Linear Probability Model observations with a negative probability or a probability exceeding 100% were dropped. The largest difference between the main Probit specification and the LPM and Logit specifications is that in the LPM model the control variable education is insignificant. Besides this, variations in the coefficient can also be observed. However, differences in coefficients between different types of regressions are difficult to interpret.
24 Specification (4) and (5) use different ESS questions to represent attitude towards
immigration. The first asks whether the respondent thinks immigrations have a negative and positive effect on the economy, and the second asks whether the respondent would like more immigration of people with different ethnicities. These variables more directly represent a respondent’s economic attitude towards immigration on the one hand and a cultural attitude towards immigration on the other. Both variables are highly significant and confirm that both economic and cultural considerations play a role in immigration.
Table 4: Robustness Check Estimation results
Specification 1)Logit 2)LPM 3)Alt unemp 4)Eco Immi 5)Cult Immi
25 Immigration -0.0119*** (economic) (0.00215) Immigration 0.0358*** (ethnicity) (0.00862) Psuedo R2 17.16% 17.34% 17.22% 17.18% R2 13.71% Observations 6453 6564 6445 6430 6462
Notes: Table contains the estimated marginal effect of the probability of voting for a RWPP, given an increase in the value
of the explanatory variable, while holding all other explanatory variables at their mean value. The standard error of the marginal effects (adjusted for clustering on country) are reported below the marginal effects. The significance level is represented by *** for 0.001, ** for 0.01, * for 0.05.
5. Conclusion
With the rise of right-wing populism in Europe it seems like Western countries are heading into another period of anti-global backlash. The election of Donald Trump and the Brexit are the most prominent large events which exemplify this phenomenon. This has also attracted the attention of many academics who have sought to explain this trend from different perspectives. Some researchers have focused on voter characteristics (Lubbers et al., 2002; Rooduijn, 2014; and Norris and Inglehart, 2016), some have focused on civil participation (Boeri et al., 2018), and some focus on economic factors (Swank and Betz, 2003; Guiso et al., 2016; Autor et al.,2016; and Rodrik, 2018). This paper also takes the economic approach in explaining the rise of right-wing populism and distinguishes itself from other literature by considering the effect of globalization by analysing the effect of immigration,
Euroscepticism, and vulnerability to globalization.
For the main research question of the paper: What are the economic determinants of the
rise of right-wing populism in Europe? Has been answered by using data from wave 7 of the
European Social survey, which was taken in 2014. To identify right-wing populist parties the classification system by van Kessel (2015) was used. Several hypotheses have been
constructed to answer the research question.
In addition to the main hypotheses I also tested whether the Low-skill, low-tech
manufacturing variable was more significant in Western-European countries than Eastern-European countries. This was done, because according to the Stolper-Samuelson theorem in Western-European countries this effect would be expected to be more significant. The results show that this was indeed the case. However, the results were not highly significant and further research specifically into this variable should be done in order to draw firmer conclusions.
26 right-wing populist voting. In contrast to other studies, age and gender were not found to be significant factors.
Unfortunately, this research has several limitations which reduce the conclusiveness of the results. Firstly, A method similar to Guiso et al (2016), where a heckprobit model is used is perhaps more suitable for this type of research, because it takes into account the condition whether or not people choose to vote in the first place. People who suffer financially or are dissatisfied with the current political system may choose not to vote in the first place (Guiso et al., 2016). Not taking this into account could lead to inaccurate conclusions. Secondly, not all countries are included in the data, the most notable missing country is Italy and the exclusion of this country could have altered the results. Thirdly, this research could be expanded by using more ESS waves to increase the number of observations. In this case this was not done, because several questions used in this research from the rotational modules of the ESS are only available in the 7th round of the ESS. Using multiple rounds would also
allow for tracking changes over time. Fourthly, in this paper the van Kessel (2015) classification of right-wing populist parties was used. However, there is always a certain degree of subjectivity when classifying political parties. For more robust conclusions multiple classification systems could be used, such as the one constructed by Inglehart and Norris (2016).
27
References
D. Acemoglu, , G. Egorov, and K, Sonin (2013). A Political Theory of Populism. Quarterly
Journal of Economics, 128 (2): 771-805.
Algan, Y., Guriev, S., Papaioannou, E., Passari, E. (2018). The European trust ciris and the rise of populism. European Bank for Reconstruction and Development, Working Paper No. 208
Autor, D. H., Dorn., Hanson., G., Majlesi, K. (2016). Importing Polarization? The Electoral Consequences of Rising Trade Exposure, NBER Working paper series No. 2637
Autor, D. H., Dorn., Hanson., G. (2018). The China Shock: Learning from Labor-Market Adjustments to Large Changes in Trade. Annual Review of Economics, 8(1): 205-240 Boeri, T., Mishra, P., Papageorigou, C., Spilimbergo, A. (2018). Populism and Civil Society.
IMF Working Paper 18/245.
Cavaille, C., Ferwerda, J. (2017. How Distributional conflict over Public Spending Drives Support for Anti-immigrant Parties. Unpublished paper, No.328
Colantone, I., Stanig, P. (2018). Global Competition and Brexit. American Political Science
Review, 112(2): 201-2018
Dustmann, C., Eichengreen, B., Otten, S., Tabellini, G., Zoega, G. (2017). Europe’s Trust Deficit Causes And Remedies. Centre for Economic Policy Research
Eurostat (2006). Nace Rev. 2 Statistical classification of economic activities in the European Community.
Available at: https://ec.europa.eu/eurostat/documents/3859598/5902521/KS-RA-07-015-EN.PDF (last access 17 June 2019)
Gereffi, G. (2014). Global Value Chains in a Post-Washington Consensus World. Review of
International Political Economy, 21(1): 9-37
Golder, M. (2016). Far Right Parties in Europe. Annual Review of Political Science, 19(1): 477-497
Guiso, L., Herrera, H., Morelli, M., Sonno, T. (2017). Demand and supply of populism. CEPR Discussion Paper DP11871.
Hatton, T. (2016). Immigration, public opinion, and the recession in Europe. Economic Policy, 31: 205–246.
have-28 nots and cultural backlash. Harvard Kennedy School Faculty Research Working Paper Series RWP16-026
van Kessel, S. (2015). Populist Parties in Europe – Agents of Discontent? Palgrave Macmillan Kitschelt H. (2002) Popular Dissatisfaction with Democracy: Populism and Party Systems.
Palgrave Macmillan
Knigge, P. (1998). The ecological correlates of right-wing extremism in Western Europe,
European Journal of Political Research, 34: 249-279.
Lubbers, M., Mérove, G., Scheepers, P. (2002), “Extreme right-wing voting in Western Europe, European Journal of Political Research, 41: 345-378.
Mayda, A.M., and Rodrik. D., (2005). Why Are Some People (and Countries) More Protectionist than Others? European Economic Review, 49(6) : 1393–1430 Mudde, C. (2007). Populist radical right parties in Europe. Cambridge University Press
Norris, P. (2005). Radical Right: Voters and Parties in the Electoral Market.
Cambridge University Press
Rodrik, D. (2018). Populism and the Economics of Globalization. Journal of International
Business Policy, 1(1): 12–33
Rooduijn, M. (2018). What Unites the Voter Bases of Populist Parties? Comparing the Electorates of 15 Populist Parties. European Political Science Review, 10 (3): 351–68 Sascha O. B., Fetzer, T. Novy, D. (2017). Who Voted for Brexit? Economic Policy 32(92): 601-
651
Schmidt, C. Alexander W., C, Spies. (2016). Immigration and Welfare Support in Germany.
American Sociological Review 81(2): 242–61
Schumacher, G,, Kersbergen, K. (2014). Do Mainstream Parties Adapt to the Welfare Chauvinism of Populist Parties? Party Politics 22 (3): 300–312
Sitter, N. (2001). The politics of opposition and European integration in Scandinavia: Is Euro‐ scepticism a government‐opposition dynamic? West European Politics, 24(4): 22-39 Swank, D. Betz, H. (2003). Globalization, the Welfare State and Right-Wing Populism in
Western Europe. Socio-Economic Review 1(2): 215–45
Timmer, M., Los, B., Stehrer, R., de Vries. G.J. (2013). Fragmentation, incomes, and jobs: an analysis of European competitiveness, Economic Policy 28 (76): 613-661.
UNESCO (2012). International Standard Classification of Education: ISCED 2011. Unesco
29 Available at: http://uis.unesco.org/en/topic/international-standard-classification-education-isced (last access 17 June 2019)
Williamson, J. G. (2002). “Winners and Losers over Two Centuries of Globalization.” NBER
30
Appendix
Appendix A: Country dummies of main specification
Specification 1)Economic 2)Cultural 3)Total 4)West 5)East
31 Poland 0.0308*** 0.0516*** 0.0468*** 0.0288*** (0.00504) (0.00658) (0.00596) (0.00238) Portugal 0 0 0 (.) (.) (.) Sweden -0.0376*** -0.0267** -0.0231* -0.00434 (0.00806) (0.00899) (0.00942) (0.00775) Slovenia -0.0282*** -0.0159*** -0.0170*** -0.0315* (0.00238) (0.00436) (0.00402) (0.0128)
Notes: Table contains the estimated marginal effect of the probability of voting for a RWPP, given an increase in the value
32 Appendix B: Correlation table of all variables
