Public political tolerance of the far right in contemporary Western Europe

Abstract Political initiatives promoting a far-right agenda have gained significant political influence in Western European democracies. This has occurred despite apparent broad-based public rejection of Europe’s Nazi past. This is a puzzle, since there are affinities between the old and the new far right. This article addresses that puzzle. Theoretically, the article distinguishes between broad and narrow interpretations of what it means to reject Europe’s Nazi past. Empirically, it shows how a well-established survey experimental template reveals substantive variations in public political tolerance of the far right. For citizens in five key Western European democracies, rejecting the Nazi past only means rejecting initiatives explicitly identified as neo-Nazi. For other far-right initiatives, political tolerance is more common and increases in accordance with these initiatives’ institutionalisation in the party system. For far-right parties fully institutionalised in the party system, public political tolerance is at the same level as for other political parties.

Ford 2020), but it is not the only one. In particular, the UN's convention on refugees, border control policies in the EU, the religious freedom of Europeans who are Muslim, and more generally the principle of free movement within the EU has been put under pressure by the rise of the far right (e.g. Givens and Case 2014;Ivarsflaten and Sniderman 2022).
Accounts of the ideology and voting for the far right emphasise the centrality of its stance against immigration, or more broadly nativism or ethno-centrism (de Lange 2007;Ivarsflaten 2008;Ivarsflaten et al. 2019;Mudde 2007;Rydgren 2005;van der Brug et al. 2005). That noted, research has repeatedly found a mixed pattern of success and failure at the polls for various far-right parties, which appears to be weakly related to the size of the demand in the electorate for their nativist ideology (Blinder et al. 2013;van der Brug et al. 2005). Notably, far-right initiatives with explicit organisational ties to the fascist milieus of the 1930s do not appear to have enjoyed increased electoral success (Art 2011;Carter 2005;Golder 2003).
The lack of success by far-right initiatives with an explicitly fascist legacy has commonly been interpreted as a result of public rejection in contemporary Western European democracies of Europe's Nazi or fascist past (e.g. Art 2007;Ignazi 1992Ignazi , 2003. However, if this is the case, then why have far-right initiatives, which often promote ideological narratives containing elements reminiscent of those of far-right initiatives of the 1930s, risen to political influence since the 1980s in some places (France, Austria, Denmark, Norway), the 1990s in others (Italy, Switzerland, and Flanders), and after 2000 in yet others (the Netherlands, Sweden, Finland, Great Britain and Germany)?
In this article we address that puzzle. We argue that the key to solving it is to improve the theoretical explanation for patterns of public political tolerance of the far right. More specifically, we need to refine two analytical concepts. First, we need a better account of which aspect of far-right ideology voters do not tolerate. We will argue that, across Western European democracies, the crucial distinction is that between broad and narrow interpretations of what it means to reject Europe's Nazi past. Secondly, we highlight the importance of analytically separating ideological features from party system institutionalisation, i.e. to what extent the far-right initiative has become institutionalised into the political system. 2 In the current literature, the two are often conceptually entangled in the concept 'new' , which can refer both to novel ideological features and to organisations that are weakly institutionalised in the system. We will argue that degree of institutionalisation into the party system, because it can serve as a cue suggesting respect for the rules of democracy, increases public political tolerance of far-right initiatives.

Public rejection of the Nazi past
One of the earliest and most persistent arguments in the scholarly literature on the contemporary far right in Western Europe holds that explicit and specific ideological links to historical fascism is a liability to mobilising a significant share of voters in elections. This was the core argument advanced by Piero Ignazi in an early agenda setting article (1992) and a later book (2003). According to Ignazi, 'the class of extreme right parties is divided into two types, depending on whether or not they are linked to fascist ideology ' (2003: 33-34). In this account, the new extreme right parties that 'deny any lineage with historic fascism' (Ignazi 2003: 32) have enjoyed increased support in elections. 3 An influential study of aggregate voting patterns operationalised this distinction between the old fascist extreme right and the new extreme right and corroborated Ignazi's claim that the latter, not the former, had increased its share of the vote in Western European democracies in the 1980s and 1990s (Golder 2003, Figure 1: 444).
In comparative analysis of individual-level voting patterns in a range of countries, Carter (2005) showed that the parties she classified as 'old extreme right' enjoyed very limited support while those she classified as 'the new extreme right' generally experienced considerable gains. Ivarsflaten (2006) additionally argued that far-right parties with a reputational shield in the form of foundational commitment to other policy-areas than immigration stood a much better chance of advancing electorally than those without such shields. Focussing on far-right party activists rather than voters, Art (2011) argued that parties with fewer personnel links to the old fascist milieus were more likely to grow and become politically influential than parties with strong such ties.

Narrow and broad interpretations
What all of these contributions found and argued in various ways was that the far-right parties that tended to be most electorally successful had managed to put some distance between themselves and historical fascism. They appeared in some way or another to be disconnected from that past and therefore in ideological terms fresh or 'new' . Here, we agree with the previous literature that the nature and extent of ideological links to fascism are important for explaining the electoral fates and developmental trajectories of nativist movements and parties in post-World War II Western Europe. However, conceiving of contemporary nativist movements as either 'new' or 'old' -and these links as either present or absent -does not bring clarity to the question of what drives public political tolerance of the contemporary far right in Western Europe.
Instead of trying to decipher whether or not far-right initiatives are 'new' or 'old' , we propose to distinguish between initiatives that would be rejected based on either a narrow or broad interpretation of what it means to reject the Nazi past. This is in line with the analyses of fascist ideologies of Griffin (1993Griffin ( , 2009), among others. He has noted that, based on a narrow interpretation of what it means to reject Europe's Nazi past, contemporary far-right parties that are commonly labelled 'new' may be right when they themselves claim that they have nothing to do with historical fascism. However, he further argues that on a broader interpretation of ideological affinity there is a link in that, '[t]heir axiomatic rejection of multiculturalism, their longing for 'purity' , their nostalgia for a mythical world of racial homogeneity (…) represent a reformist version of the same basic myth' (Griffin 2000: 174).
Such ideological analyses highlight continuities and similarities in the narratives and structure of ideas in both so-called new and old far-right initiatives. A larger body of scholarly literature acknowledges that the question of which political initiatives should be tolerated in contemporary liberal democracies and which should not, is both normatively tricky and politicised (Bleich 2011;Capoccia 2005;Eatwell and Goodwin 2018;Mudde 2019). Our proposed distinction between narrow and broad interpretations acknowledges and follows from these important ongoing scholarly debates.
A similar distinction is Mudde's (2007) separation between the 'radical' right and the 'extreme' right. In Mudde's usage the radical right is associated with anti-pluralism, while the extreme right is associated with being explicitly anti-democratic. The distinction we introduce between narrow and broad interpretations of what it means to reject Europe's Nazi past overlaps with Mudde's distinction between radical and extreme right in the sense that far-right organisations rejected on a narrow view are likely explicitly anti-democratic. That said, the narrow interpretation criterion is both more concrete and more restrictive than Mudde's 'extreme right' concept. This means that there could be parties that would be classified as extreme according to the definition proposed by Mudde that nevertheless would not be rejected on a very narrow interpretation of what it means to reject the Nazi past.

Institutionalisation into the party system
The institutionalisation of political parties has commonly been referred to as an organisational process, using measures of stability and solidity (Huntington 1968;Panebianco 1988). This study departs from that view of party institutionalisation. Instead of looking at party age, number of candidates or electoral support, we examine the process of becoming institutionalised into the party system, or more specifically far-right organisations' level of incorporation in the political system.
When the existing literature on the far right distinguishes between the old and the new far right, the questions of institutionalisation and ideological affinity to the Nazi past are often blurred. The theoretical framework we propose analytically disentangles these separable dimensions. One lesson to be learned from the history of democratic break-downs in Europe is that system institutionalisation is no guarantee against authoritarian takeovers and/or democratic break-downs (i.e. Capoccia 2005; Levitsky and Ziblatt 2018;Linz and Stepan 1978). Loewenstein recognised this in 1937, when he warned that democracy could become the 'Trojan horse by which the enemy enters the city ' (1937: 424).
Institutions such as governments, can be a source of information to voters and signal what norms are acceptable or not (Tankard and Paluck 2016). Norris (2005) identified early on that political institutions could extend legitimacy to far-right initiatives; '(…) radical right parties have been surging in popularity in many nations, gaining legislative seats, enjoying the legitimacy endowed by ministerial office, and entering the corridors of government power' (Norris 2005: 3). More recent research has argued that parliamentary representation of far-right political parties contributes to normalisation and legitimacy of the far right (Valentim 2021). Valentim further notes that: 'Importantly, the centrality of parliaments in democratic systems (…) means that the signal about normative change provided by parliamentary representation is highly public' (Valentim 2021(Valentim : 2482. The framework we propose extends and further systematises these insights. We rely on the well-established scholarship on stages in the life-cycle of political parties to conceptualise the thresholds of institutionalisation into the party system relevant to our analysis. In the foundational study, Pedersen (1982) identified four such thresholds. We focus on three of them. 4 The 'threshold of declaration' is when an organisation publicly declares its intent to run for election and thus in many definitions becomes a political party rather than an extra-parliamentary movement, interest organisation, or advocacy group. According to Pedersen, 'the declaration of intent is usually coupled with some overt acts like campaign, collecting of signature for registration, various media events, etc.' (Pedersen 1982, 6). Importantly, this threshold is not related to the size of the support base, as smaller organisations may well declare their intent to run for elections, while larger organisations may choose not to do so.
'The threshold of representation' delineates political parties that run for elections but fail to receive enough votes to be elected to representative bodies from those that succeed in transferring votes into seats. In Pedersen's (1982: 7) own words, the threshold of representation is 'the barrier which all parties have to cross in order to obtain seats in the legislature' . In European politics, there are of course examples of parties that receive representation on the local or EU levels yet are not represented in national parliaments. 5 A more fine-grained analysis could incorporate these intermediary thresholds, but in this account we focus on representation in national parliaments.
The final stage identified by Pedersen is the threshold of 'relevance' , which he connects to Sartori's work on the coalition or blackmail potential of political parties (Sartori 1976: 122-123). In this study, we put the bar a bit higher than that and focus instead on the threshold of government participation (Deschouwer 2008). Parties that have been included in government have a different set of experiences than parties that have not, and, importantly for the question at hand, the public has had an opportunity to observe the party's behaviour when in power.

Far-right movements in five Western European countries
The proposed two-dimensional theoretical framework for the study of political tolerance of the far right is summarised in Figure 1. The first dimension concerns ideology and whether rejection of that ideology would be based on a narrow or broad interpretation of what it means to reject the Nazi past. The second dimension concerns institutionalisation into the party system with the four phases delineated by the thresholds in the Pedersen framework. As these are complex issues, our theoretical framework does not propose that these are the only factors affecting public political tolerance of the far right. However, our novel contribution is to propose that these two dimensions jointly account for the main pattern of public political tolerance of far-right initiatives.
We use the theoretical framework to examine public political tolerance of a variety of far-right initiatives in five Western European countries: France, Germany, the Netherlands, Sweden, and Norway. The question of public political tolerance of the far right is highly relevant in all of these countries. The initiatives to be examined were deliberately and carefully selected to secure variation on both explanatory dimensions: ideology and institutionalisation. We were to some extent constrained by reality, since we wanted to examine public political tolerance of real political initiatives whenever feasible. Figure 1 shows how we place all the far-right initiatives included in the study in the proposed theoretical framework. To ensure that we had one example of a far-right initiative indisputably connected to the Nazi past in the narrowest possible interpretation, we in all countries study public political tolerance of organisations labelled neo-Nazi. All the other far-right initiatives included in the study would be rejected only on a broader interpretation of what it means to reject the Nazi past. This latter bundle of initiatives is otherwise highly heterogeneous, including both anti-Islamic extra-parliamentary organisations and a variety of well-known and much studied far-right political parties. They were selected to ensure variation with regards to institutionalisation in the party system. Below, we will discuss the placement of the far-right initiatives in the theoretical framework.
Since specific extra-parliamentary organisations are not widely known among the public, we examine public political tolerance of them using generic labels (neo-Nazi; anti-Islamic). Extra-parliamentary initiatives such as PEGIDA (Patriotische Europäer gegen die Islamisierung des Abendlandes) make up a sizeable, and predominantly transatlantic, far-right anti-Islamic movement. These initiatives advocate a version of far-right ideology that sets them apart from historical Nazism and fascism along some ideological lines. For instance, many espouse support for the Jewish minority while drawing on liberal values, such as gender equality and LGBT-rights, to define who they are and represent (e.g. Berntzen 2020a; Fangen 2020).
That said, two core elements of their ideology clearly fit the criteria for 'generic fascism' . First, they portray Islam and Muslims as an existential enemy set to destroy Western civilisation. Second, they believe that this external enemy is enabled and abetted by an inner enemy -a gullible and traitorous political elite variously described as 'politically correct' , 'multiculturalists' , 'globalists' , or 'cultural Marxists' (e.g. Berntzen and Sandberg 2014). Therefore, if public rejection of the far right is based on a broad interpretation of what it means to reject the Nazi past, public political tolerance of anti-Islamic initiatives would be low. If, instead as we have proposed, public political tolerance follows from a narrow interpretation, the public will be considerably more likely to tolerate anti-Islamic than neo-Nazi organisations.
We expect the far-right initiatives that have crossed the threshold of declaration and therefore by most definitions have become far-right parties to be well known to the public in all five countries under study. We therefore use their proper names in the questionnaire. They vary with respect to extent of their institutionalisation in the party system. At the time of the study (June 2017), Die Alternative für Deutschland (AfD, Germany) had participated in elections but was not represented in national parliament or government. Others had been elected to national parliament, including Front National (FN, France), 6 Sverigedemokraterna (SD, Sweden) and Partij voor de Vrijheid (PVV, the Netherlands), while yet others were part of government coalitions, including Fremskrittspartiet (FrP; Norway).
The FN and the SD have histories of explicit ties with Nazi or neo-Nazi organisations/personnel, ideologies, and symbols. The SD was founded in 1988 and grew partly from an organisation that embraced Nazi symbols and ideas. Still, due to the reforms made after Jimmie Åkesson became leader of the party in 2005 (Rydgren and van der Meiden 2019), at the time of the study (2017), the SD would be rejected only on a broader interpretation of what it means to reject the Nazi past.
The Front National was founded in 1972, when various extreme-right groups came together to form a united alliance (Rydgren 2005). The leader of the party for decades, Jean-Marie Le Pen became infamous throughout Europe for insisting, despite convictions of hate crime, on his right to refer to the Holocaust as 'just a detail in the history of World War II' (Snipes and Mudde 2020: 453). This was a prominent way in which the party's leader maintained a direct link to the specifics of Nazi ideology. However, in this case more recent developments are also significant. Marine Le Pen replaced her father as party leader in 2011. She excluded him from the party in an explicit attempt to de-demonize and reform the party (e.g. Ivaldi 2016). At the time of this study, rejection of the French National Front would therefore require a broader interpretation of what it means to reject the Nazi past.
For the other far-right parties included in the study, the argument for placing them in the 'broad interpretation' category is more straight-forward. The PVV was founded in February 2006 by Geert Wilders. The PVV subscribes broadly to the ideological narrative described above for anti-Islamic organisations (Vossen 2017). Wilders has argued for banning the Koran, expelling Muslims from Europe, and introducing a tax on headscarves (Vossen 2011, 185). Both due to these policies and his efforts to support other anti-Islamic initiatives such as PEGIDA in Germany, Wilders has been characterised as a key figure in the broader anti-Islamic movement. Wilders has been prosecuted for hate speech and discrimination several times (van Spanje and de Vreese 2015;Verkuyten 2013), and anti-Islamic activists have consistently responded by staging demonstrations supporting him (Berntzen 2020a: 75).
The AfD was founded in 2013 as a single-issue party criticising the Euro, but it quickly turned against immigrants and Muslims (Schmitt-Beck 2017). The AfD is a particularly interesting case because it is the first far-right party with major national electoral success in Germany after the Second World War. Reminiscent of discussions about new far-right parties elsewhere in Europe, there has been significant debate in Germany about whether the AfD should be considered an extreme right party (Arzheimer 2019). In Germany, such discussion has special significance, as extreme right parties are banned by the constitution. After the AfD became the third largest party with 12.6 percent of the vote in the national election in the fall of 2017, the controversial figure and AfD politician Jens Maier called it the 'biggest win since 1945' (Siri 2018: 141). Most would agree, however, that because of the AfD's efforts to distance itself from the specificities of Nazi ideology, organisation, and symbolism, when the study was conducted in the spring of 2017 it would only be rejected based on a broad interpretation of what it means to reject the Nazi past.
The final case in the framework, The Norwegian Progress Party (Fremskrittspartiet, FrP) is another example of a party where scholars have disagreed about whether it should be classified as far-right (e.g. Jupskås 2015). The party was founded in 1973 as a protest party against increased taxes, fees, and bureaucracy (originally called Anders Langes parti). Beginning in the late 1980s, the nativist agenda became a prominent pillar of the party's ideology. In 2013, the party entered government for the first time as a minority partner in a Conservative-led minority government. 7 It was in government at the time of the study and surrounded by controversy. Much of this controversy originated in the party's past and contemporaneous promotion of core anti-Islamic ideas, arguing for instance that the Labour Party was responsible for the 'sneak Islamization' of Norway (Berntzen 2020b;Jupskås 2015).
We have placed the far-right parties and movements in their respective cells in Figure 1 based on our theoretical framework. It is important to note that this framework is a simplification. As elaborated above, the parties have a variety of historical experiences and backgrounds. For instance, the FN and SD have historical ties to fascist organisations and ideology, while the PVV has not. Still, they are all placed in the same cell in our theoretical framework, because none of them at the time of the study explicitly identified as neo-Nazi. They therefore would be rejected only on a broad interpretation of what it means to reject the Nazi past.
The theoretical framework separates ideology from institutionalisation. It does so, because the question of whether becoming institutionalised in the party system requires ideological moderation is central and unresolved in the study of the far right. We know that party system institutionalisation can occur with ideological moderation. The Sweden Democrats is a case in point. They crossed the threshold of representation only after significant revisions of their party program and a change in leadership (Rydgren and van der Meiden 2019). What remains debated is the extent to which institutionalisation into the party system can contribute to public political tolerance absent ideological moderation.

Study design
To generate the data needed to examine patterns of public political tolerance of the far right, we conducted the same survey experiment in five countries. We deliberately did not invent a novel design but instead adapted one of the classic survey experimental templates that has been successfully employed in the study of tolerance (Petersen et al. 2011;Stouffer 1955;Sullivan et al. 1979). Our novel contribution was to randomly vary which far-right initiative voters were asked about in a between-subjects design. 8 The full wording of the survey experiment was as follows: Imagine that […group] has asked to rent a community center to host a meeting for its members and sympathizers. How much do you agree or disagree that […group] should be allowed to rent the center for this purpose?
This question was highly relevant at the time and in the contexts under study. In contemporary Europe, far-right initiatives have been denied access to renting a local community house on several occasions. One well-known example from our country case selection is the Sweden Democrats, who in the end were able to rent a location for their election night gathering in the 2014 general election, but only after misrepresenting the true purpose of the gathering. When the owners of the location found out on election night that the SD was using the space for their main election night event, the SD suffered a black-out (Palm 2014).
An important feature of the experimental design is that in each country we ask one of the randomly selected subsets of the sample about a political party fully institutionalised in the party system that is not on the far right but rather an established, small centre-right party. 9 Responses to these centre-right parties provide a baseline indicator of maximum political tolerance of non-far-right parties with small voter bases that are fully institutionalised in the party system. Public political tolerance of the far-right initiatives included in the study will be compared and contrasted to this baseline.

Data
The experiment was conducted as a part of the European Internet Panel Study (EIPS). The EIPS was a pilot of a collaborative internet panel survey that facilitated coordinated comparative survey experiments (Arnesen 2018). 10 The experimental data collected in five key Western European countries was administered in the following online panels: The Norwegian Citizen Panel (Ivarsflaten 2017) The five panels are all high-quality probability based online panels with representative samples. This EIPS 2017 joint survey included questions on themes such as climate change, migration, diversity, and populism. 11

Results
We begin the analysis by establishing the benchmark results. The expectation is that responses to the question about ordinary, small, centre-right parties will be a good indicator of maximum public political tolerance of a party fully institutionalised in the party system. The results are displayed in Figure 2. In all five countries, the centre-right parties are granted the democratic right to assemble by nearly all respondents. Less than 10 percent of the sample oppose the idea that they should be allowed to rent the local community house. The benchmark indicator of maximal political tolerance is thus successfully established. To make the figures more easily readable, the benchmark indicator is merged for all five countries in further analyses. 12 Next, we turn to the first results of substantive interest: public political tolerance of neo-Nazi organisations. They were included as indisputable examples of organisations that are likely to be rejected by citizens based on a narrow interpretation of what it means to reject Europe's Nazi past.
Previous research leads us to expect broad public rejection. Figure 2 shows that this is precisely what we find. In all five countries we find that neo-Nazi organisations are not tolerated by the vast share of citizens, between 70 and 90 percent. This result confirms in a comparative survey-experimental setting what previous research has suggested -the public in contemporary Western Europe rejects far-right initiatives that directly and explicitly link themselves to the ideologies, labels, or symbols of Europe's Nazi past. Moreover, the results showing political tolerance towards a neo-Nazi group confirm that this is a unanimous pattern of rejection that does not differ between countries with different experiences with the Nazi past.

Broader interpretations
We now turn to the results for the far-right initiatives that, like the neo-Nazi organisations, are extra-parliamentary, but unlike them they have made efforts to decouple themselves from the specificities of historical Nazi and fascist ideology and symbolism: the anti-Islamic organisations. A comparison of public political tolerance of neo-Nazi and anti-Islamic organisations is ideal for observing whether citizens reject far-right initiatives consistent with a broad or a narrow interpretation of what it means to reject Europe's Nazi past because they are in the same phase of institutionalisation into the party system -they are undeclared extra-parliamentary organisations. Figure 3 displays the results. In all countries, public political tolerance of anti-Islamic organisations falls somewhere between the broadly rejected neo-Nazi organisations and the widely tolerated centre-right parties. In other words, the public appears divided on the question of whether extra-parliamentary organisations with ideological affinities but no explicit links to the specificities of historical Nazism should be tolerated. In other words, while the rejection of far-right initiatives based on a narrow interpretation of rejecting the Nazi past is overwhelming and near consensual, rejection based on a broader interpretation is contested within the public. Figure 3 reveals not one, but two sets of patterns that have not been previously seen. First, there are clear, substantively large, and statistically significant differences between public political tolerance of anti-Islamic organisations and the fully tolerated centre-right political parties. The size of this difference is about 30 percentage points. A substantial share of citizens would reject the extra-parliamentary anti-Islamic organisations. But, just as importantly, the contrast between public political tolerance of the anti-Islamic and the neo-Nazi organisations is of similar magnitude. Again, the difference is about 30 percentage points. A vast majority of citizens, on average about 80 percent, reject the neo-Nazi organisation, consistent with a narrow interpretation of what it means to reject the Nazi past. Only about half of all citizens reject the anti-Islamic organisation, consistent with a broad interpretation of what it means to reject the Nazi past.

Institutionalisation in the party system
The results so far have shown that the distinction between narrow and broad interpretations of what it means to reject Europe's Nazi past is an important source of variation in public political tolerance of the far right in contemporary Western Europe. Moreover, we found a large gap between the benchmark indicator of the centre-right political parties that are fully institutionalised in the party system and the political tolerance of extra-parliamentary anti-Islamic organisations.
This gap could result from differences in the levels of institutionalisation in the party system, as suggested in the theoretical discussion above. If these levels influence public political tolerance of the far right, then we should observe differences not only according to the narrow and broad interpretations of what it means to reject the Nazi past but also according to the level of party system institutionalisation. In contrast, if institutionalisation in the party system is unimportant, then we should observe a pattern influenced only by the distinction between narrow and broad interpretations. More specifically, if this is the case, then public tolerance of far-right initiatives that have crossed one or more of the institutionalisation thresholds should line up around the level of anti-Islamic organisations. Figure 4 below shows the results for far-right initiatives that have crossed one or more of the institutionalisation thresholds. A first major take-away is that there is considerable variation in public tolerance that cannot be explained by the distinction between broad and narrow interpretations. In fact, public political tolerance of these far-right initiatives varies across a large spectrum, from the FrP, which is fully tolerated, to the AfD, which is tolerated only by about half of German citizens. Public political tolerance of far-right initiatives has not been examined in this way before, but we can still conclude that the scale of the observed variation is surprisingly large.
In Figure 4, the stark difference in public political tolerance of the PVV compared to the anti-Islamic organisations is especially informative. As laid out in detail above, the PVV embraces an anti-Islamic ideology very similar to the one promoted by the extra-parliamentary anti-Islamic organisations. This is the case where we may most confidently state that the initiative differs only with respect to the extent of institutionalisation into the party system. If the level of institutionalisation had not influenced the political tolerance of far-right organisations, then the levels of public political tolerance of the PVV should be the same as that of the anti-Islamic activist organisations. However, as seen in Figure 4, this is far from the case. Only a small minority of Dutch citizens, about 15 percent, reject the political party, PVV, which is close to fully institutionalised in the party system. This compares to as many as about 50 percent of a different random subset of the same sample of Dutch citizens rejecting anti-Islamic organisations. The difference in the levels of public political tolerance is about 35 percentage points, which is roughly the same order of magnitude as the impact of the distinction between broad and narrow interpretations of what it means to reject the Nazi past. This is strong evidence, even if not exact causal identification, to suggest that party system institutionalisation of far-right initiatives matters for public political tolerance.
Another result in Figure 4 is also striking but needs to be interpreted with more caution. Political tolerance of the far-right party in Norway (FrP) is indistinguishable from that of the centre-right benchmark. This party is the only far-right initiative included in this study that receives a full score on the institutionalisation dimension. We do not know whether this high level of political tolerance will persist after the party exits government or if it is contingent on being in government at the time of the study. The interpretation of results is further complicated by the fact that the FrP is also more ideologically moderate and diversified than the anti-Islamic extra-parliamentary organisations. Nevertheless, the result is consistent with the argument that institutionalisation in the party system impacts public political tolerance of the far right.
Another important result also shown in Figure 4, and consistent with the proposed theoretical framework, is that the least tolerated of the far-right initiatives that has entered the electoral arena (AfD) is also the political party that is least institutionalised into the party system. At the time this study was conducted, the AfD was on course to stand for national election, where it did gain representation. However, when a sample of citizens in Germany responded to the survey, the AfD did not yet have a seat in national parliament. A follow-up study is needed to determine whether crossing the threshold of representation increased the level of public political tolerance of the AfD.
One reason to believe it would increase the level of tolerance is on display in Figure 4. The level of public political tolerance for the far-right parties that had crossed the threshold of representation at the time of study (SD; FN) is considerably higher than that seen for the AfD. The relevant figures are approximately 30 percent who reject the SD and FN compared to approximately 60 percent who reject the AfD. Indeed, the similarity of the results for the SD and the FN are consistent with the institutionalisation hypothesis, as both parties are at a similar stage in Pedersen's life-cycle model. They are represented in parliament but have never been included in government.
The pattern of results in Figure 4 are thus highly consistent with the proposed theoretical framework, which highlights both ideological affinity with historical Nazism and the extent of institutionalisation in the party system. Far-right organisations outside of the electoral arena with explicit links to the specific ideological ideas, labels, and symbols of historical Nazism are rejected by the public. Other far-right initiatives, where the ideological affinities with the European far-right of the 1930s are of a more general nature, are much more likely to be tolerated. Any gaps that such far-right initiatives experience in political tolerance compared to the benchmark close as they become institutionalised in the party system.
Finally, there are some potential limitations and remaining questions regarding the account we have proposed of public political tolerance of the far right in contemporary Western European democracies. Most notably, public political tolerance of the AfD, which at the time had crossed the declaration threshold and hence had become a political party, is at approximately the same level as that of extra-parliamentary anti-Islamic organisations. This result could mean two things. First, it could be an indication that the threshold of declaration is less important than we had hypothesised. If so, far-right initiatives only receive a public tolerance premium after being elected to national parliament and not merely by declaring the intention to stand for election. Since the AfD is the only example in this study of a far-right initiative that had only crossed the declaration threshold, we should refrain from concluding decisively based on the available data.
A second possible explanation for this pattern has to do with different national cultures of remembrance (Caramani and Manucci 2019). Perhaps the historical lesson that the Nazis came to power through elections and the way this historical lesson has been handled by schools and other public institutions in Germany has led to broader awareness that party system institutionalisation is an unreliable indicator of democratic intent and respect for democratic institutions (e.g. Art 2011). Future studies could address the question of German exceptionalism empirically. Note that the results we have presented here caution against overemphasising the role of different national cultures, since the main patterns we find can be accounted for by the general two-dimensional framework we have proposed. Capoccia (2005) wrote the following about the classic liberal dilemma of tolerating the intolerant: '(…) on the one hand, the need for tolerance and, on the other, the danger that the intolerant represent a threat to the persistence of the very principle of tolerance, has historically been solved in practical terms' (Capoccia 2005: 5). The results from this study show that, for ordinary citizens, the tolerance of the intolerant has boundaries, but they are drawn quite narrowly. We find that voters fully reject far-right organisations only on a very narrow interpretation of what it means to reject the Nazi past.

Discussion and conclusion
In this study we have argued and shown that two dimensions strongly affect public political tolerance of the far right in contemporary Western Europe: (1) whether the given initiative is directly and explicitly linked to the specific organisations and ideological content of historical fascism and Nazism (narrow interpretation of affinity to Europe's Nazi past) and (2) the extent to which the far-right initiative has become institutionalised into the party system.
Using this theoretical framework, we were able to design studies that revealed empirical patterns of tolerance of far-right initiatives that, to the best of our knowledge, have not been seen before. We learned that consensual public rejection of the far right is fairly uncommon and appears to be the rule only for extra-parliamentary organisations with affinities to historical Nazism and fascism based on a narrow interpretation of what that means. For far-right initiatives that are tied to fascism based on a broader interpretation of affinity, there is disagreement among citizens in all the five democracies we studied. The public division is starkest in the case of extra-parliamentary initiatives, i.e. the anti-Islamic organisations. For far-right initiatives that have become institutionalised in the party system, public political tolerance tends to be more common than not. For the far-right initiative in our study that had crossed all thresholds of institutionalisation, we observed no penalty in public political tolerance at all compared to the centre-right benchmark party.
This broad set of results holds in all five Western European democracies in our study with only some minor exceptions. They are consistent with those of a number of previous studies, which have argued that explicit ideological links with historical fascism is detrimental to the contemporary far right (e.g. Art 2011;Carter 2005;Ignazi 1992Ignazi , 2003. The new contribution here is clarifying that the public consensus on rejection of the old far right is based on a very narrow interpretation of what that means. There is no such public consensus for rejecting far-right initiatives based on broader interpretations. On top of this central clarification, an additional contribution of this study is to underscore the importance of becoming institutionalised in the party system for public political tolerance of the far right. The most striking new aspect of the evidence we provided, is the considerable extent of variation in public political tolerance of the far right in contemporary European democracies. Not least, public political tolerance of far-right initiatives that would be rejected based on a broad interpretation varies from highly contested (50/50 splits) to uniformly tolerated. As we pointed out, the difference in political tolerance of the PVV, which is an anti-Islamic far-right party that has crossed the threshold of representation, and an anti-Islamic organisation that has not crossed the threshold of declaration is stark. The ideology is the same, but the extent of institutionalisation is different.
Fielding the Far-Right Tolerance experiment simultaneously in five countries generated new data on variation in public political tolerance. In particular, the far-right organisations included in the experiments offered a good range of variation when it comes to one of the central analytical concepts advanced in this study, that of institutionalisation in the party system. That said, there are limitations to the conclusions that may be drawn especially from the part of the experiment where we measure tolerance of specific (and well-known) far-right parties, such as the PVV and the AfD. We can be absolutely confident that the PVV had crossed the threshold of representation at the time of the study and therefore was more institutionalised into the party system than was the AfD, which had only crossed the threshold of declaration. We can also be absolutely confident that the share of the public that tolerated the PVV, was much larger than the share of the public that tolerated the AfD. So the new observations generated by the study presented here are consistent with the institutionalisation hypothesis. We have however not established that institutionalisation in the party system caused the increase in public political tolerance. This is because for all the real-life far-right parties included in the study, there are a number of other additional characteristics that also differ between them, and we cannot rule out that these matter.
More studies are needed to rule out other alternative hypotheses. This study brings the theoretical and empirical scholarly debate about tolerance of the far right a step further by offering a parsimonious and plausible analytical account and showing new data, collected in a manner not seen before in the study of the far right, which is consistent with this theoretical account. More research is needed to disentangle some of the questions not settled by the current study. The AfD in Germany and the SD in Sweden stand out as particularly interesting cases to examine further because they have, or are about to, cross additional thresholds of institutionalisation. The AfD is now represented in parliament, and the SD may become included in the governing coalition, at least as a supporting party, after the Swedish 2022 national election.
Moreover far-right parties with stronger ties to the Nazi past should be included in future studies, examples could be the NPD in Germany, the NVU in the Netherlands, the BNP in Great Britain, or the Brothers of Italy. The results reported here lead to the hypothesis that even more ideologically extreme parties, and parties with various forms of ideological or organisational roots in Europe's Nazi past, would be tolerated at least on a par with anti-Islamic organisations unless they themselves identify as neo-Nazi and there was widespread agreement in society that they were neo-Nazi. Future research should further test the generalisability of this conjecture. This analysis and these results provide important new insights to the study of the far right in political science. The finding that a large share of European citizens generally employs a very narrow definition of what it means to reject the Nazi past when they decide which far-right organisations to tolerate means that the societal agreement to reject the Nazi past does not extend to a consensus on rejecting anti-Islamic organisations. We know of no other study demonstrating this with similar levels of control of potentially confounding factors.
The finding that institutionalisation in the party system appears to impact public political tolerance of the far right can be a cause for concern. To the extent that party system institutionalisation does not produce ideological moderation and/or commitment to democratic rules and norms, the pattern we have observed is worrisome. We have shown that organisations promoting ideologies that about half of all citizens would hesitate to tolerate if they were promoted by extra-parliamentary organisations are tolerated by all or nearly all if they receive enough votes to become normalised through representation in parliament. This 'democracy premium' can in turn render attempts at disciplining these actors ineffectual or counterproductive, as the public believes they should be tolerated.

Notes
1. In this article we use the label 'far right' when referring to the broad set of political initiatives -parties and non-parties -other studies have referred to some or all of these initiatives in a number of different ways, including the populist radical right, the extreme right, the far right, the alternative right, and neo-fascist. 2. The concept of party system institutionalization has mostly been used to describe the stability of interparty competition and the stability of party roots in society (e.g. Mainwaring and Scully 1995). The concept we use is related but different when we refer to the process of institutionalization into the party system and not the end-state of the party system as a whole (whether it is an institutionalized system or not). 3. Most initiatives that Ignazi and others then characterized as the new extreme right are commonly described as the populist radical right today, whereas some would still be described as extreme right. 4. Pedersen also identified a threshold of authorization in between declaration and representation. While this distinction certainly may be important for some questions in party research, we believe that it has less relevance for the problem at hand here. 5. It is a well-documented fact that second-order elections benefit fringe-parties compared to national elections (e.g. Reif and Schmitt 1980). 6. Now renamed Rassemblement National. 7. The Christian People's Party (Kristelig Folkeparti) and the Liberal Party (Venstre) were invited to join the government but decided only to cooperate as supporters of the government and not to formally join due to their ideological distance to the Progress Party. They eventually entered the coalition government during its second term, joining in January 2018 and 2019, respectively. 8. More information on the wording and coding of the dependent variable can be found in the online appendix.
9. The specific center-right parties included in the experiment was: Venstre in Norway, Freie Demokratische Partei (FDP) in Germany, Democraten66 in the Netherlands, Les Republicains in France, and Liberalerna in Sweden. 10. Iceland also collaborated but is not included in this study due to the lack of any established far right initiatives. 11. For more information on the individual panels, sampling, and distribution, see Arnesen (2018). 12. All results are presented with their marginal means and are based on ANOVA models. Full model outputs are presented in Tables 4-6 in the online appendix.