Introduction
Representative democracy in general, and traditional political parties in particular, is in crisis (Mair, Reference Mair2013; Ignazi, Reference Ignazi2017). Although this is nothing new (Ignazi, Reference Ignazi1996; Gunther et al., Reference Gunther, Montero and Linz2002), the truth is that support for anti-political-establishment parties (APEp) is on the rise, especially since the beginning of the 2008 Great Recession (Norris and Inglehart, Reference Norris and Inglehart2019; Kriesi and Pappas, Reference Kriesi and Pappas2015; Wolinetz and Zaslove, Reference Wolinetz and Zaslove2018), and has never been so high. Whereas, as shown elsewhere (Rama and Casal Bértoa, Reference Rama and Casal Bértoa2019: 11), after the Second World War the average level of electoral support for APEp was just 11%, it almost doubled during the last decades.Footnote 1 For example, only during the last years, APEp such as Alternative for Germany (AfD), the National Front, or the League (Lega) as well as the Five Stars Movement (M5S) have managed to shake the foundations of other traditionally consolidated democracies in Germany, France, and Italy, respectively.
APEp have even managed to get access to government in long-established Western European democracies such as Denmark or Norway. In most cases, they have done so as simple junior coalition partners (e.g., Austria, Belgium, Finland, and, most recently, Spain). In other cases (just two), they have managed not only to win parliamentary elections but also to form totally anti-political-establishment coalition governments: the first one in Greece, where the Coalition of the Radical Left (Syriza) joined forces with the national conservative, but also with Eurosceptic, Independent Greeks (ANEL) in 2015, and the second one, more recently (2018), in Italy between the populist M5S and the regionalist (also populist) Lega.
The recent electoral success of APEp in general, and far-right parties in particular, has raised important questions about the future of democracy in Western Europe (Mudde, Reference Mudde2019; Rama and Casal Bértoa, Reference Rama and Casal Bértoa2019). In fact, the rise of support for APEp (e.g., Communists, Fascists, and populists) has been traditionally associated with executive fragility (Powell, Reference Powell1981), mass rioting (Lane and Ersson, Reference Lane and Ersson1994), systemic instability (Sartori, Reference Sartori1976), democratic backsliding (Enyedi, Reference Enyedi2016), and, finally, democratic breakdown (Linz, Reference Linz1978). In a similar vein, Foa and Mounk (Reference Foa and Mounk2016) as well as Galston (Reference Galston2018) also pointed out how the recent success of populist/extreme parties (another type of APEp) has a negative effect on (liberal) democracy.
Given this strong and negative correlation between rise in support for APEp and democracy, especially its liberal (i.e., rule of law, minority rights and separation of powers) component (Mudde and Rovira Kaltwasser, Reference Mudde and Rovira Kaltwasser2018: 4), it is essential we try to understand the main factors behind APEp’s success. Drawing on a new comparative dataset including all 20 Western European democracies since 1849,Footnote 2 the current article attempts to provide an explanation to the rise in electoral support for APEp. In particular, the article examines three alternative explanations traditionally put forward by the literature, namely, economic, institutional, and sociological. The main aim of the article is to see to what extent they are in contradiction or, on the contrary, complementary to each other.
The article proceeds as follows. The following section discusses the conceptualization and operationalization of the dependent variable and shows the evolution of support for APEp by country and over time. The third section looks at the literature on the determinants of support for APEp – be they populist, radical, extreme and anti-systemic – and formulates various hypotheses. The fourth section presents the main independent variables as well as the various controls. The dataset and the methodology are introduced in the fifth section, while the sixth section presents the results. The article concludes with a summary of our main findings, containing also some reflections on what their implications might be for the future development of Western European party politics.
APEp in Western Europe
The concept of APEp is not universally accepted. Many scholars prefer to employ the term ‘radical’(Funke et al., Reference Funke, Schularick and Trebesch2016), ‘populist’ (Norris and Inglehart, Reference Norris and Inglehart2019), ‘anti-system’ (Zulianello, Reference Zulianello2019), ‘pariah’ (van Spanje and van der Brug, Reference Van Spanje and Van Der Brug2007), ‘extreme’ (Carter, Reference Carter2005), ‘challenger’ (Hobolt and Tilley, Reference Hobolt and Tilley2016), or ‘protest’ (Morlino and Raniolo, Reference Morlino and Raniolo2017) to refer to political parties sharing a clear anti-establishment character.
Trying to avoid this nominological confusion (Zulianello, Reference Zulianello2019), and in order to capture all those parties that pose a challenge to democracy (independently of their populist character or not, their location – at the extremes or not – within the political spectrum, the way – discriminative or not – establishment parties treat them, etc.), we have decided to use the ‘more comprehensive’ (Schedler, Reference Schedler1996: 292) notion of APEp. In particular, and following Abedi’s original definition, we consider that an APEp needs to present at the same time all of the following three characteristics:
(1) to perceive itself as a challenger to the parties that make up the political establishment; (2) to assert that a fundamental divide exists between the political establishment and the people (implying that all establishment parties, be they in government or in opposition, are essentially the same); and (3) to challenge the status quo in terms of major policy issues and political system issues (Reference Abedi2004: 12).Footnote 3
In practice, however, our dataset includes most of the parties labeled by scholars as radical, populist, extreme, anti-immigration, and protest, as many of them also clearly fulfill all of Abedi’s three criteria at the same time.Footnote 4
A complete list of all the APEpFootnote 5 for the elections included in our analyses can be found in Table A1 in the Online Appendix. For the 1945–2000 period, we accept Abedi’s original list (Reference Abedi2004: 143–150) as a valid roster of APEp. For the period after 2000, we rely on multiple sources (including both case- and medium-N studies), but especially on the following four works: Loomes (Reference Loomes2011), van Kessel (Reference Van Kessel2015), Kriesi and Pappas (Reference Kriesi and Pappas2015), and Wolinetz and Zaslove (Reference Wolinetz and Zaslove2018). For the period before 1945, and on top of multiple country-based works/datasets (e.g., Hanley, Reference Hanley2002; Karvonen, Reference Karvonen1993; Vossen, Reference Vossen2003), we mostly based our coding on Berg-Schlosser and Mitchell (Reference Berg-Schlosser and Mitchell2000), Capoccia (Reference Capoccia2005) and Linz (Reference Linz1978).
Figure 1 shows the percentage of votes for APEp (our dependent variable) in 20 Western European democracies since the beginning of the 20th century.Footnote 6 In agreement with the most recent literature, electoral support for this set of parties is on the rise (Norris and Inglehart, Reference Norris and Inglehart2019; Hernández and Kriesi, Reference Hernández and Kriesi2016; Wolinetz and Zaslove, Reference Wolinetz and Zaslove2018). As it follows from the figure above, this is clearly the case all-over Western Europe, with just two exceptions,: namely, Malta and Switzerland.Footnote 7
Figure 1. Percentage of votes for APEp in Western Europe, 1900–2017.
Furthermore, and following the conventional wisdom, most of the countries where APEp have been more successful are those affected by, among other factors, (1) economic downturn (e.g., Cyprus, Greece, and Iceland); (2) immigration (e.g., Germany, France, and The Netherlands); (3) corruption (e.g., Spain); or all these three at the same time (e.g., Italy).
On average, the percentage of votes for APEp in Western Europe during the last few years has exponentially increased: from around 17% during the previous two decades (1990–2009) to more than 24% in the 2010s (2010–2017). To the point that more than the 40% of the elections with record levels of support for APEp held since 1950 have taken place only during the last decade. In fact, Western European APEp attract now 13% more votes than in the 1960s, when party systems were considered to be ‘frozen’ (Lipset and Rokkan, Reference Lipset and Rokkan1967).
However, it is also important to note that in an important number of countries (i.e., Germany, Finland, France, Ireland, Italy, Norway, Switzerland, Portugal, and Spain), the election with the highest percentage of votes for APEp took place before the beginning of the 21st century. Moreover, it is also possible to observe in Figure 1 that in most consolidated European democracies the current levels of electoral support for APEp are much lower than during, for instance, the interwar period (1918–1939).Footnote 8
Given the cross-national and cross-temporal variation observed here, and before trying to answer the question on the sources of APEp’s support in Western Europe so we can learn how to tackle the specter that is currently haunting this part of the old continent, we will proceed now to revise the factors that, according to the existing literature, could explain the success of this set of parties.
Economic, institutional, sociological change, and electoral support for APEp
Economic factors
The rise of support for APEp experienced during the last decade, and especially in those countries mostly affected by the 2008 Great Recession (e.g., Greece, Spain), has favored the view that economic conditions are mostly responsible for the high levels of support this set of parties has enjoyed in many Western European democracies. Based on the literature that shows how voters reward parties in government when economic conditions are favorable and punish them when the economy performs poorly boosting the opposition (Stegmaier and Lewis-Beck, Reference Stegmair and Lewis-Beck2000; Dassonneville and Lewis-Beck, Reference Dassonneville and Lewis-Beck2019), the recent scholarship has suggested a clear positive relationship between negative macroeconomic factors and support for APEp.
In a first large-N comparative analysis of the causes of electoral support for extreme-right parties in Western Europe between 1970 and 1990, Jackman and Volpert (Reference Jackman and Volpert1996) found that electoral support for this kind of party increases in those countries where the electoral system is proportional and unemployment is high. In a follow-up work including also the 1990s, Golder (Reference Golder2003) found support for the latter, but not the former. Interestingly enough, he also found that high levels of immigration also boost electoral support for extreme-right parties.Footnote 9 The positive relationship between high levels of unemployment (contextual factor), as well as being unemployed (individual factor), and electoral support for the extreme right was also found in Arzheimer’s (Reference Arzheimer2009) multi-level study of 18 Western European countries between 1980 and 2002. These findings were clearly in line with Abedi’s (Reference Abedi2004) work which, using the broader notion of APEp and including both extreme-right and extreme-left parties, found a similar positive relationship between unemployment and support for non-mainstream parties.Footnote 10
More recently, Doležalová’s (Reference Doležalová2015) study of the levels of electoral support for extremist (both left and right) parties in 23 EU member states between 1995 and 2012 also found unemployment to be a major cause. However, and echoing Brückner and Grüner’s (Reference Brückner and Grüner2010)Footnote 11 finding of a moderate positive relationship between economic slowdown and support for extreme political forces in Organization for Economic Co-operation and Development countries, they also showed an association between economic recession and the rise of extremism.
Bearing in mind all that has been said, we expect support for APEp to increase in an economic hardship context.Footnote 12 The main reason is that under negative economic conditions voters will be more inclined to support these kinds of parties proposing alternative solutions to their sufferings. In this regard, and following Kayser and Wlezien (Reference Kayser and Wlezien2011), we measure economic negative circumstances by using the levels of Gross Domestic Product (GDP) growth. Thus, we hypothesize that
HYPOTHESIS 1a: Economic decline and APEp’s electoral support are positively related.
However, and given the large amount of recent works linking the 2008 global financial and economic crisis with the surge of APEp in the last few years (Lindvall, Reference Lindvall2014; Roberts, Reference Roberts2017), it is reasonable to think that what triggers voters to support APEp is not so much the state of the economy per se, but the shock caused by so-called ‘big crisis’ (e.g., Great Depression and Great Recession) as they combine both financial and economic hardships with cultural and political downturns. To the point that, as it has been shown elsewhere (Achen and Bartels, 2016), voters’ punishment might be determined by particular uncontrollable events, for instance, shark attacks, droughts, or floods. In this context, it is also plausible to think that the kind of shocks caused by severe (and global) crises – like the Great Depression or the Great Recession – might lead stressed citizens to behave ‘irrationally’ and, consequently, support an alternative party. Also, as some scholars have recently pointed out (Fernández-García and Garcia-Luengo, Reference Fernández-García and García-Luengo2019; Rico and Anduiza, Reference Rico and Anduiza2019; Mols and Jetten, Reference Mols and Jetten2016, Reference Mols and Jetten2017), it might be the perception that the country is doing poorly economically – rather than the actual economic decline – that make voters feel attracted toward the ‘easy solutions to complex problems’ proposed by most APEp. For example, recent studies such as the one conducted by Mutz (Reference Mutz2018) find that support for Donald Trump in the US 2016 presidential elections could be explained by increasing anxiety among high-status groups, rather than pocketbook economic concerns.
In fact, in what can be considered the most comprehensive study of the impact economic and financial crises have on the level of support for extreme-right parties, Funke et al. found that ‘policy uncertainty rises strongly after financial crises as government majorities shrink and polarization rises’ (Reference Funke, Schularick and Trebesch2016: 1), especially after the crisis, but often lasting less than 8 years. More recently, but looking at a different period (1900–2016), Dalio et al. (Reference Dalio, Kryger, Rogers and Gardner2017) also suggested a clear correlation between economic downturns and support for populist parties.
In a similar vein, but adopting a case-study approach which looked at the sources of populism in several European countries, Kriesi and Pappas (Reference Kriesi and Pappas2015) concluded that it was the 2008 economic crisis that led to an increase in the level of support for populist (both left and right) parties. This argument was later confirmed by Kriesi in a work together with Hernández (Reference Hernández and Kriesi2016) where they also pointed out that parallel to the decrease of support for established political parties, economic hardship mostly benefited the radical-left, the radical-right, but also other non-mainstream (APEp) parties (e.g., the Greens).
Hernández and Kriesi (Reference Hernández and Kriesi2016) clearly show that, unlike what happens under normal economic circumstances, in elections taking place under the context of the 2008 Great Recession, economic voting theories lose explanatory power in favor of more long-term (systemic change) approaches. In their view, big economic crises do accelerate pre-existing de-alignment (as voters perceive that established parties do not represent their interests) and re-alignment (as new social divides gain relevance in order to determine vote choices) processes. Thus,
in contexts like the Great Recession, punishment of the incumbents by the voters is not only likely to occur in much greater proportions, but its consequences are also likely to be longer-lasting than the shifts of the voters to the mainstream opposition parties (Reference Hernández and Kriesi2016: 204).
The idea, therefore, is that under particular critical economic junctures, like the ones in 1929 or 2008, both economic and political resentment will benefit parties fighting the status quo (Hooghe and Marks, Reference Hooghe and Marks2018; Roberts, Reference Roberts2017). In other words, and paraphrasing Casal Bértoa and Weber (Reference Casal Bértoa and Weber2019: 234), ‘big crises’ are like the ‘treatment’ that causes a ‘response’ in domestic party politics, namely, the rise of APEp. Thus, and taking into consideration what has been said, we can formulate the following hypothesis:
HYPOTHESIS 1b: During big economic events, the impact of economic variables on support for APEp will be higher than in other time periods.
Institutional factors
Notwithstanding what has been said, and although the idea of a positive relationship between economic downturns and support for APEp sounds intuitive, the fact that this type of party obtained excellent electoral results in economically growing countries like Poland (Casal Bértoa and Weber, Reference Casal Bértoa and Weber2019: XXVII-XXX), France, The Netherlands, Austria, or Finland (EEAG, 2017: 60) lead us to think of other possible (non-economic) explanations (Mudde, Reference Mudde2007: 206).
Basically, one can distinguish two types of these ‘non-economic’ explanations in the literature, namely, institutional and sociological. Posed by the ‘cartel party thesis’ school, the first type puts the blame on traditional political parties as they are considered to have become, especially since the late 1980s, ‘state agencies’ failing to represent voters’ interests (van Biezen, Reference Van Biezen2004).
As it follows from Katz and Mair’s (Reference Katz and Mair1995) seminal work, these two types have been the main reasons why the ‘cartelization’ of traditional (e.g., conservative, social-democratic, and liberal) political parties has led to an increase in support for APEp. First of all, the ‘collusion’ of mainstream parties, and consequent convergence toward the center of the political spectrum, has made it very difficult for voters to (ideologically) distinguish among them, making important segments of the electorate more susceptible to be attracted to the markedly different policies put forward by APEp. Secondly, the abovementioned move of mainstream parties to the center has also left an important (unrepresented) space at the margins of the ideological spectrum, prone to be occupied by those APEp (Abedi, Reference Abedi2002; Mair, Reference Mair and Kerman2002; Morgan, Reference Morgan2013). In other words,
where moderate left and right parties have converged toward centrist positions and may even have cooperated in government coalitions, the chances for ‘populist antistatist parties’ as well as parties of the ‘New Radical Right’ [or Left] to be electorally successful rise considerably’ (Kitschelt and McGann, Reference Kitschelt and McGann1995: 17, 20–23, 48).
The result has been the appearance of many new parties which, claiming to defend the interests of the people against the ‘corrupted’ elites (Mudde and Rovira Kaltwasser, Reference Mudde and Rovira Kaltwasser2017), have come to occupy such ‘representative void’, generally found at the fringes of the political spectrum (Katz and Mair, Reference Katz and Mair2009).Footnote 13 Consequently, we hypothesize that
HYPOTHESIS 2: The higher the electoral fragmentation, the higher the support for APEp.
If, as a result of the process of cartelization by which traditional political parties have withdrawn from society into the state losing its traditional representative function and becoming ‘public utilities’ (van Biezen, Reference Van Biezen2004), new parties (the majority of them characterized by an anti-establishment rhetoric) have emerged progressively gaining ground (both in the electorate and in government) at the expense of mainstream ones, then electoral fragmentation could be considered as a good proxy to measure the extent to which traditional (mainstream) political parties are in crisis.
Sociological factors
Another school of thought, led by Dalton and other American scholars, puts the emphasis instead on sociological – rather than institutional – change (Müller-Rommel, Reference Müller-Rommel, Müller-Rommel and Bértoa2016: 4–5). The basic idea behind this line of thought is that the process of globalization and secularization, common to all Western democracies since the late 1970s and especially after the end of the Cold War, has brought important changes in society which, in turn, have led to a decline in what Sartori (Reference Sartori1969) called ‘the sociology of politics’. In other words, contrary to what – at least until the late 1980s – had been the norm in Europe, where the structure of inter-party competition in a country was determined by its cleavage structure (Lipset and Rokkan, Reference Lipset and Rokkan1967; Bartolini and Mair, Reference Bartolini and Mair1990), the decline of – mainly the class and religious – cleavages has led to a process where the appearance of new issues (e.g., migration and corruption) and dimensions of competition (e.g., austerity and Europeanization), less socially rooted, have produced a decrease in party identification, an increased importance of candidate and performance-evaluation and/or valence voting, and an important reduction in electoral stability (Dalton and Wattenberg, Reference Dalton and Wattenberg2000; Dalton and Weldon, Reference Dalton and Weldon2005; Thomassen, Reference Thomassen2005).
At the aggregate level, the classic way to measure such decline in the sociological roots of voters’ partisan preferences, mostly caused by the declining electoral relevance of traditional cleavage groups, has been Pedersen’s index of total electoral volatility (TEV) (see Crewe, Reference Crewe, Crewe and Denver1985; Franklin et al., Reference Franklin, Mackie and Valen1992), traditionally seen also as an indicator of ‘electoral realignment’ (Pedersen, Reference Pedersen1979: 6). For all these reasons, we think that
HYPOTHESIS 3: TEV and support for APEp are positively related.
Institutional and sociological factors
However, more recently, in what has been labeled as a ‘Copernican Revolution’ (Casal Bértoa, Reference Casal Bértoa2017: 305) or a ‘Damascene convergence’ (Dalton et al., Reference Dalton, Farrell and McAllister2011: 24), Mair came in his posthumous study (Reference Mair2013) to the conclusion that the crisis of representative democracy and the consequent rise of APEp had its cause on the ‘mutual withdrawal’ of both political parties and voters from the electoral arena.
Thus, and contrary to what could be observed in the so-called ‘golden age’ of ‘mass party’ democracy, where both voters and parties were interconnected by a ‘representative’ chain that went from voters – whose ideological inclinations were anchored in social divisions – to the state through parties, acting as the agents which glued ideological preferences with government policies; now not only voters have deserted political parties as data on declining party membership, trust, and voter turnout clearly show, but also parties have abandoned voters, taking refuge in the state. A simple look at the levels of party financial dependence, over-regulation, and patronage would be more than illustrative (van Biezen and Kopecký, Reference Van Biezen and Kopecký2014; Mair, Reference Mair2013). As a result, and taking into consideration that not only voters but also parties should be blamed for the current wave of APEp’s success, we formulate a fourth hypothesis combining both schools of thought (i.e., institutional-European and socio-American):
HYPOTHESIS 4a: The higher the electoral fragmentation and volatility, the higher the vote for APEp will be.
In addition, and following the big events argument explained above, we could expect that such an incremental effect of GDP growth in the support for APEp during this time period could expand not just to economic variables, but also to institutional and sociological ones. The argument is the same: the two big economic crises (i.e., Great Depression and Great Recession) will amplify the effect of both the crisis of parties and the crisis of voters.
HYPOTHESIS 4b: During the Great Depression and the Great Recession, we expect a higher and positive impact of electoral fragmentation and volatility on APEp support.
Economic, institutional, and sociological factors: operationalization
Following the scholarship reviewed in the previous section, we will proceed now to operationalize each of the independent variables. However, and before examining each of them in turn, it is important to note that because we are aware that some of our hypotheses could have a different direction (e.g., the higher the support for APEp, the higher the electoral fragmentation in a party system), and in order to avoid any endogeneity problems, we opted for lagging all the independent variables, not just the economic ones. Doing so, we address the impact that economic, institutional, and sociological factors measured in ‘t-1’ (in the previous election year) might have on the levels of support for APEp in time ‘t’. Table A2 in the Online Appendix presents descriptive statistics (number of observations, mean, standard deviation, minimum, and maximum) for all of the variables mentioned in this article.
In order to test for the effect of the economy on the level of support for APEp, we use GDP growth. This is not just due to data availability, but also – and especially – to the central role this indicator has traditionally had when measuring economic performance (Kayser and Wlezien, Reference Kayser and Wlezien2011). Specifically, for each election, we use the level of GDP growth (lagged). GDP data come from Gapminder (2017).Footnote 14
Following Sartori’s (Reference Sartori1976) classic work, linking the number of parties with the presence of powerful anti-systemic/APEp, we will use Laakso and Taagepera’s (Reference Laakso and Taagepera1979) standard ‘effective number of electoral parties’ (ENEP) to capture the extent of the crisis of traditional political parties. Measuring how many parties are in a given election weighted according to size,Footnote 15 this indicator constitutes an excellent proxy to try to measure the extent to which new parties have made a successful breakthrough in a given party system (Taagepera and Shugart, Reference Taagepera and Shugart1989).
Electoral change, or the abandonment of traditional political parties by voters, is measured here by Pedersen’s (Reference Pedersen1979) classic index of TEV, which captures ‘the net change within the electoral party system resulting from individual vote transfers’.Footnote 16 Given that the recent success of APEp has coincided in time with an increase in the level of ‘extra-system’ volatility in many European countries (Chiaramonte and Emanuele, Reference Chiaramonte and Emanuele2017), high TEV scores are assumed to show the extent to which voters are swinging their preferences from traditional to new – ‘anti-political-establishment’ in most cases – parties.
All in all, in our models, ENEP attempts to capture the crisis of representative democracy from the parties’ points of view, while TEV looks at the voters’ perspective. In both cases, the data come from Casal Bértoa (Reference Casal Bértoa2020).
As we also expect a positive relationship between electoral fragmentation and support for APEp, we control for the disproportionality of the electoral system, calculated on the basis of Gallagher´s (Reference Gallagher1991) ‘least square index’. The idea is that because more proportional electoral systems will allow more parties into the party system (Duverger, Reference Duverger1954; Lijphart, Reference Lijphart1994), including those with an anti-political-establishment character, the percentage of votes for APEp will be higher the less the disproportionality of the electoral system (see also Fernández-García and Garcia-Luengo, Reference Fernández-García and García-Luengo2019).Footnote 17
Another institutional variable for which we control is that of type of regime (TOR). Ever since the publication of Linz’s seminal work on the perils of presidentialism (Reference Linz1990a) and the virtues of parliamentarism (Reference Linz1990b), non-parliamentary regimes have been considered to have negative implications for the healthy functioning of democracy, including increasing the level of support for anti-systemic parties. The idea is that while in parliamentary regimes the head of state is either hereditary or (s)elected either by a super-powerful government, which already assembles the qualified majority usually required for this type of (s)election, or by a compromise with the opposition; in (semi-)presidential regimes the ‘double electoral process’ gives APEp the possibility to win office or, at least, get the necessary publicity to express their grievances against the establishment. The most recent 2017 French presidential elections, which constituted a boost for Le Pen’s Front National or Melenchon’s Unbowed France, constitute a clear example.Footnote 18 We operationalize parliamentary regimes as 1 and (semi-)presidential regimes as 0.Footnote 19
We also control for (other) sociological factors. Using Alesina et al.’s (Reference Alesina, Devleeschauwer, Easterly, Kurlat and Wacziarg2003) classic dataset, we introduce time invariant variables that capture the level of sociological fractionalization presented in a particular country, namely, linguistic, ethnic, and religious (Lipset and Rokkan, Reference Lipset and Rokkan1967). The logic behind this control is double. On the one hand, scholars have shown again and again the positive relationship between sociological fractionalization and electoral fragmentation (Cox, Reference Cox1997; Geys, Reference Geys2006) as well as the negative association between the former and TEV (Bartolini and Mair, Reference Bartolini and Mair1990; Bernhard and Karakoc, Reference Bernhard and Karakoc2011). On the other hand, we expect higher sociological diversity to foster higher support for APEp in a direct way, given the nativist and anti-immigration character of most of these parties (Boomgaarden and Vliegenthart, Reference Boomgaarden and Vliegenthart2007).Footnote 20
Finally, we also control for the democratic experience a country has and, in particular, for the time (i.e., years) a democracy has lasted (YoD). While, following Converse (Reference Converse1969), one could expect more consolidated democracies to be vaccinated against anti-political-establishment discourses threatening the democratic system in comparison to new ones, where traditional parties and voters are assumed to have experienced the ‘APEp’s challenge’ less often; the ‘cartel party thesis’ explained above, according to which party systems in long-standing consolidated democracies have become over-institutionalized, could lead us to think exactly the contrary.Footnote 21
Data and method
The dataset that we use to test our hypotheses comprises Western European party systems under democratic conditions. We consider a country to be democratic when (1) it has a score of ≥6 in the Polity IV index, (2) universal (male) suffrage elections have been held at least once, and (3) governments are formed with (and rely on) parliamentary support, rather than on the exclusive will of the head of state (Casal Bértoa, Reference Casal Bértoa2020). Some countries, such as France or Finland, comprise different periods according to the Republic in question. Other countries, such as Austria or Germany, refer to two different party systems (1st and 2nd Republic or the Weimar and Bonn Republics). Some democracies had to be excluded given the lack of data on electoral results.Footnote 22 The end result is a new dataset with 20 Western European democracies, comprising more than 470 elections between 1849 and 2017 (see Table 1).
Table 1. Time period and number of elections for the selected countries
Source: Casal Bértoa (Reference Casal Bértoa2020).
Because this is an unbalanced time-series cross-section dataset (Beck and Katz, Reference Beck and Katz1995), containing an average of 24 elections by country (minimum 10 for Cyprus and maximum 34 for Switzerland), we make use of panel data with elections nested in countries, which limits substantially the set of methods for the analysis. In order to solve these problems, we estimate an ordinary least square (OLS) with robust standard errors within the country clusters in order to prevent problems of heteroscedasticity and autocorrelation (see Emanuele et al., Reference Emanuele, Chiaramonte and Soare2018). The model estimated is the following one:
$$\displaylines{
APEp\;support = \alpha + \beta \left( {GDP\;growth} \right) + \beta \left( {ENEP} \right) + \beta \left( {TEV} \right) + \beta \left( {Lsq} \right) + \beta \left( {TOR} \right) \cr
+ \beta \left( {Religiously} \right) + \beta \left( {Linguistic} \right) + \beta \left( {YoD} \right)e \cr} $$Results
Figure 2 displays the OLS results for our main modelsFootnote 23 (see also Table A3 in the Online Appendix).Footnote 24 The graph shows the standardized (standard deviation = 1) independent variables. As we have mentioned, we run various OLSs with country clusters – to address the hierarchical structure of the data – and robust standard errors. Model 1 (see also model 1, Table A3 in the Online Appendix) presents the effect of GDP growth on the electoral success of APEp. Although, following the most recent scholarship (Doležalová, Reference Doležalová2015; Brückner and Grüner, Reference Brückner and Grüner2010), we had expected poor economic conditions to foster electoral gains for APEp (see Hypothesis 1a), but looking at the figure below, we can clearly reject this hypothesis.
Figure 2. Explaining APEp’s support in 20 Western European democracies (1849–2017).
Interestingly enough, none of our control variables (models 3, 4, 5, and 6) seems to be significantly related to our dependent variable with one exception: TOR. The latter is moderately (only at 95%) and negatively related to our dependent variable, meaning that support for APEp is higher in (semi-) presidential than in parliamentary regimes, endorsing Linz’s fears regarding the perils of direct presidential elections. All other variables – even if not significant – have the expected sign, confirming that the percentage of vote for APEp is higher in countries with (1) higher religious and linguistic diversity (positive sign), (2) higher disproportionality (negative sign), and (3) younger democracies (see the signs of the coefficients in model 6, the full model, in Table A3).
Conversely, both ENEP (traditional parties’ crisis) and TEV (voter’s change) have positive and very significant impact on the level of electoral support for APEp (see Models 2 and 6), confirming our Hypotheses 2 and 3. This clearly suggests that contrary to the literature that opposes institutional and sociological factors both have contributed to APEp’s electoral success. In other words, rather than contradictory, both (European and American) schools of thought are complimentary, as both the crisis of traditional political parties and electoral change are to be blamed for the rise in support for APEp. This is more clearly illustrated in Figure 3, which shows the linear prediction of support for APEp as a function of fragmentation (left) and volatility (right). In particular, every time that an ‘effective’ electoral party enters the party system, the percentage of votes for APEp will increase in more than 3.5 points. In a similar vein, every time the TEV rises one point the vote for APEp will experience an increase of 2% (see also models 2 and 6 in Table A3 in the Online Appendix). All in all, Figure 3 clearly supports our Hypothesis 4a.
Figure 3. Linear prediction of the percentage of vote for APEp as a function of electoral fragmentation (ENEP) and total electoral volatility (TEV).
Following Lindvall (Reference Lindvall2014), as well as evidence from Figure 1, which shows how support for APEp increased in many European countries mostly after 1929 (e.g., Belgium, Greece, Germany, and Spain) and 2008 (e.g., Denmark, Greece, Iceland, and Spain), but not so much after 1973 (Italy and Denmark are perhaps the only exceptions), we will then limit our comparison to the first two (e.g., Great Depression and Great Recession) crises,Footnote 25 operationalized as dummies in Figure 4.
Figure 4. The effect of big economic events on the support to APEp.
We test Hypothesis 1b in models 4 and 5. Model 1 (replication of model 6 in Figure 2) constitutes the base model. Thus, Figure 4 shows that whereas the Great Depression dummy does not have an effect (neither positive or negative) on the support for APEp (model 2), the Great Recession dummy (model 3) exhibits a high (statistically significant) coefficient (i.e., 11.6; see Table A4 in the Online Appendix). This clearly shows that, in the context of the 2008 Great Recession, voters were more willing to support APEp. In a similar vein, the interaction among GDP growth and both ‘big economic events dummies’ only results as statistically significant in the case of the Great Recession (model 5). The sign is negative and the coefficient is extremely high (i.e., −16), suggesting that the lower the GDP growth in Western European democracies since 2008, the higher the share of votes for APEp. Trying to better illustrate the result of both interactions, Figure 5 displays the linear predictions.
Figure 5. Interaction between big economic events (1929 and 2008) and GDP growth.
Thus, and contrary to all the literature on the effects of the 1929 ‘Great Crash’ (e.g., Klapsis, Reference Klapsis2014; de Bromhead et al., Reference De Bromhead, Eichengreen, Rourke and Kevin2012), Figure 5 (left) clearly shows that it was not the economic hardships of the Great Depression that led voters to support APEp (e.g., the Nazis in Germany or the Communists in Spain) and, consequently, helped many Western European democracies to collapse. Conversely, however, the right plot in Figure 5 partially confirms that Hypothesis 1b, as one of the big crises (i.e., the Great Recession), is significantly associated with APEp’s success. As Kriesi and Pappas stated it was ‘during the Great Recession [that] populism in Europe increased notably’ (Reference Kriesi and Pappas2015: 323), to the point that ‘is currently at its highest level since the late 1930s’ (Dalio et al., Reference Dalio, Kryger, Rogers and Gardner2017: 1).
However, if the shock of such a ‘big event’ like the Great Depression is able to amplify the impact of the economy on APEp’s support, we should expect such a big crisis also to amplify the effect of our two main explanatory factors increasing the share of votes for APEp in more fragmented and volatile party systems.
In this context, Figure 6 aims to shed some light on the relationship between our two main explanatory factors (i.e., mainstream parties’ crisis and electoral change) and support for APEp during the two big economic crises (i.e., 1929 and 2008).Footnote 26 As it follows from models 1 and 2 in Figure 6, neither fragmentation nor volatility is significantly related to APEp’s electoral success during the Great Depression. This is in a way not surprising because no matter how severe the 1929 Great Depression was, it took place at a time when both (‘mass’) parties and (especially class) cleavages were strongest (Bartolini and Mair, Reference Bartolini and Mair1990; Sartori, Reference Sartori1976).
Figure 6. Interaction between big economic events (1929/2008) and fragmentation/volatility.
The situation after 2008 is totally different. In fact, as it follows from models 3 and 4 in Figure 6, the Great Recession helped to significantly increase the impact of both electoral fragmentation and volatility on APEp’s support. This seems to confirm Rovira Kaltwasser and Zanotti’s conclusion that ‘the Great Recession has [simply] amplified pre-existing electoral trends’ (Reference Rovira Kaltwasser and Zanotti2017: 8). Not in vain, the process of party and cleavage decline, and the consequent thawing of Western European party systems, had already started various decades ago (Mair, Reference Mair1997; Sartori, Reference Sartori, Linz and Valenzuela1990). What the Great Recession has done, in contrast to what happened during the Great Depression at the height of Lipset and Rokkan’s (Reference Lipset and Rokkan1967) ‘freezing hypothesis’ and ‘mass parties’ golden age’, is to make those already weak representative democracies even more vulnerable. In already weakened party systems, both by the crisis of traditional political parties and social change, the shock of the 2008 global financial and economic crisis has lead to the boost of the decline of mainstream parties and the rise of APEp.
Figure 7, by highlighting the interaction effects of the Great Recession dummy, underlines the findings of models 3 and 4 above (see Figure 6) and confirms Hypothesis 4b for the case of the Great Recession. As it follows from the figure above, the impact of electoral fragmentation (left) on the percentage of votes for APEp was way higher after the Great Recession than before. And the same can be said in relation to TEV (right).
Figure 7. Relationship between fragmentation/volatility and APEp support during the Great Recession.
It is important to note, additionally, that when we take into consideration the whole dataset (1849–2017) fragmentation is more relevant and statistically significant than volatility (see Figure 3), but after the 2008 Great Recession the relationship is totally the opposite, that is, an increase in one TEV unit corresponds to an increase of 3.8 in the percentage of votes for APEps, whereas every time a party enters in the system, APEp support increases ‘only’ in more than 2.7 points. In other words, during the Great Recession, the impact of electoral fragmentation in APEp support has been one point lower than that of TEV. Consequently, we can conclude that at least for the period compressed between 2008 and 2017, ‘sociological change’, possibly inspired by the ‘the emergence of a transnational cleavage, which has as its core a political reaction against European integration and immigration’ (Hooghe and Marks, Reference Hooghe and Marks2018: 109), has been more responsible for the rise in support for APEp than the crisis of traditional parties themselves.
All in all, it seems clear that ‘a single model’, as the ones posed by the European (institutional) and American (sociological) schools,
is insufficient to describe the diverse changes occurring in contemporary party systems. The changes affecting these systems are too diverse, ranging from the emergence of new parties to growing anti-party sentiment (Flanagan and Dalton, Reference Flanagan and Dalton1984: 7).
More precisely, and as it follows for our analyses, only a combination of both (institutional and sociological) factors can help us to really understand the rise of APEp and their electoral success (Mair, Reference Mair2013).
Conclusions
Support for APEp is on the rise. Never since the turbulent inter-war period has APEp managed to be so successful, not only at the electoral level, but also at the governmental arena. By the end of 2017, APEp governed, either as senior or junior partners, in various Western European democracies.Footnote 27 This has led many scholars, but also journalists and practitioners, to blame the 2008 global and financial economic crisis for APEp’s electoral success.
In the current article, we look at the factors that have traditionally been associated with the rise of APEp, namely, economic, institutional, and sociological. Using an original database comprising 687 APEp and more than 400 elections between 1849 and 2017 in 20 consolidated Western European democracies, and aware of the limitations, this type of aggregate and statistical studies pose to the disentanglement of any eventual individual mechanism (i.e., the sociological and social psychological processes that fuel support for APEp), we found no relationship between economic performance per se and electoral support for APEp. Instead, our results clearly show that both institutional (electoral fragmentation) and sociological (TEV) factors are significantly associated with APEp’s electoral success. In other words, neither the crisis of political parties (European school) nor voters (American school) can be unilaterally blamed for the APEp wave, but the ‘mutual withdrawal’ of both (parties and voters) from the electoral arena. As Mair put it in his posthumous work, it has been ‘the widening gap between rulers and ruled [that] has facilitated the often strident populist challenge that is now a feature of many advanced European democracies’ (Reference Mair2013: 18–19).
Remarkably, and notwithstanding many similarities between the 1929 Great Depression and the 2008 Great Recession (Lindvall, Reference Lindvall2014), we also found that while the former did not have any significant impact on APEp’s success, the latter did. To the point that it not only amplified the positive effect economic decline traditionally has on the vote for APEp, but also clearly boosted the negative influence both party decline and social change were already having on their electoral success. In this context, our results confirm that the 2008 Great Recession, which combines economic, political, and cultural features, is really unique. In this context, the article sheds light on the literature about the causes of support for APEp in general, and populist parties in particular, by emphasizing the need to adopt a historical approach that, on the one hand, looks beyond the current post-2008 crisis and, on the other hand, goes beyond the traditional economic and cultural unilateral explanations.
Also interesting, especially from the point of view of trying to find possible solutions to the current ‘anti-establishment’ wave and consequent crisis of liberal democracy, is to point out that while before the 2008 global and financial crisis the crisis of traditional parties, with the ‘cartel style’ attitudes, had mostly led voters either to seek refuge at the arms of APEp or to abandon the political arena once and for all (Morgan, Reference Morgan2013), after the Great Recession it seems more a matter of social change and the appearance of new cleavages (Hooghe and Marks, Reference Hooghe and Marks2018).
This is certainly an important point that we need to remember when trying to find a cure to the rise of APEp. Because if increase in the support for APEp is just the symptom of a major problem, that of a democratic system is characterized by the disenchantment of citizens with traditional political parties that have not managed to adapt to a new social reality and have therefore failed to represent their new interests; then the antidote should be neither party bans nor cordon sanitaires, but a new remedy that leads to the revitalization of the traditional parties at the core of the democratic body.
Acknowledgements
We would like to especially thank Professors José Ramón Montero (Universidad Autónoma de Madrid), Cees van der Eijk (University of Nottingham), Russell J. Dalton (University of California), and Dr. Anna Lührmann (University of Gothenburg) for their comments on a previous version of this manuscript as well as the three anonymous reviewers for their valuable comments. We also want to express our gratitude to participants at the IDEA/ODIHR/NIMD/WFD/REPRESENT International Conference on ‘Representation in the Age of Populism’ held in Brussels between June 18–20, 2018; as well as to Guillermo Cordero, Andrés Santana, Simona Guerra and the rest of participants at the 2019 Council for European Studies Annual Conference held in Madrid for their helpful comments on a preliminary version of this manuscript.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S1755773920000260.
