Selective tolerance and the radical right

In recent decades, there has been increased tolerance within many countries towards a range of previous out-groups. This has been displayed most dramatically in the growing acceptance of the LGBT community. Some radical right organisations are also expressing tolerance towards the very same out-groups which they once reviled. We postulate that the radical right strategically uses tolerance to increase its own support and to impose costs on another out-group – Muslims – who are judged to be hostile to the tolerated groups. We provide a theoretical analysis and contrasting case studies that help explain the conditions under which radical right organisations will, or will not, display tolerance towards out-groups.


Introduction
In recent decades, there has been increased tolerance within many countries towards a range of groups. This is most dramatically displayed in the increasing acceptance of the LGBT community, but it can also be seen in reduced racism, anti-Semitism and sexism. Radical right political parties and organisations are also increasingly espousing tolerance towards some of these previous out-groups. Recent years have seen the 'detoxification' of France's Front National (now National Rally) by leader Marine Le Pen which has led the party to advocate for gay rights and adopt a more liberal position on abortion. At the same time, however, the Front National adopted a more hostile stance towards Islam (Mayer, 2013(Mayer, , 2015. This approach has arguably assisted the competitiveness of the party in French elections as social values appear to take on more prominence than the traditional economic left/right divide in politics (Facchini and Jaeck, 2019). In the Netherlands, openly-gay Pim Fortuyn led a populist movement calling for an end to Muslim immigration before his assassination in 2002 and the political space he occupied has since been filled by Geert Wilders' Party for Freedom, which similarly combines tolerance towards women's rights, LGBT people and Jews with hostility towards Islam (Akkerman, 2005;Vossen, 2011). In the United Kingdom, the radical right, anti-Islam street protest organisation, the English Defence League (EDL), has LGBT, Sikh and Jewish divisions, and rainbow and Israeli flags have been displayed at its marches as protestors launch hostile, abusive chants at Muslims (Meadowcroft and Morrow, 2017;Pilkington, 2016). Increasingly, it seems that radical right parties and organisations in North-western Europe are voicing their support for liberal principles while simultaneously rejecting Muslims on the grounds of their allegedly illiberal values (see Simonsen and Bonikowski, 2020).
This change of attitude towards groups who were previously a target of radical right organisations is usually understood to be a strategic decision aimed at increasing and retaining popular support. It is argued that publicly supporting the rights of women, Jews and LGBT people enables the radical right to signal its rejection of biological racism and fascism and thereby broaden its appeal beyond the small number of people comfortable with more extreme views. Numerous scholars have noted that radical right organisations that have distanced themselves from biological racism and fascism have successfully mobilised significant numbers of people to vote or take to the streets in their support, whereas those that have not taken steps towards detoxification have failed to achieve similar success (Golder, 2003;Goodwin, 2011;Mayer, 2015).
A stereotypical radical right group would ideally choose to express intolerance towards a range of out-groups as this would cement a narrow and clearly defined group identity. In Western countries, that identity could be White, predominantly male, heterosexual and Christian. However, the group also wishes to maintain sufficient membership to be socially and politically relevant. It may be that in some societies intolerant views are not stigmatised, and in these societies, one should expect to see broadly intolerant radical right groups. However, in societies where sexism, racism and homophobia are stigmatised, the adoption of a more tolerant stance that reduces the costs of stigma will help to maintain and expand membership. At the same time, such an approach risks undermining the very raison d'etre of a radical right group; if tolerance is extended too broadly, the existential purpose of a radical right organisation ceases. This is because the organisation would no longer be able to uphold the xenophobia and nationalism that, along with the punitive moralism of authoritarianism, render it 'radical right' (Mudde, 2007). Therefore, the twin tasks of maintaining and extending membership, while also maintaining the intolerance that the group requires, is achieved by focusing attention primarily on one out-group and currently this is the Muslim community. Reducing stigma through tolerance is perceived to impose costs on the Muslim community which is judged to be more traditionally conservative and therefore negatively affected by the broader acceptance of liberal rights. Expressions of tolerance by the radical right are invariably couched in terms linked to their perception of Muslims as intolerant. Tolerance is not expressed independently, as a stand-alone value in itself, without reference to Muslims. This strongly suggests the expressed tolerance is strategic and perhaps not wholly sincere. 1 In the next section, we provide further background on patterns of tolerance and intolerance within wider society and the radical right. We then outline a potential model for selective tolerance. We provide empirical evidence from studies of two UK-based organisations, the EDL and the United Kingdom Independence Party (UKIP), and from a study of Poland's Law and Justice Party, to support our theory. The article finishes with a brief conclusion.

Tolerance, stigma and the radical right
Tolerance may be defined as 'respect for diversity' (Corneo and Jeanne, 2009: 691) and 'openness, inclusiveness . . . to all ethnicities, races and walks of life' (Florida, 2003: 10). As Berggren and Nilsson (2013: 178) have noted, tolerance may be consistent with like and dislike, approval or disapproval, of those who are tolerated -the relevant characteristic is that diversity is tolerated, not that it is liked or meets approval.
Tolerance has been positively linked to economic growth (Andersen and Fetner, 2008). Berggren and Elinder (2012) and Nilsson (2013, 2016) have found evidence of a positive relationship between economic freedom and tolerance, particularly of homosexuality. Given the evidence of a positive relationship between economic growth and tolerance, and economic freedom and tolerance, it is unsurprising that there is evidence of a general global trend towards greater tolerance in recent decades that coincides with increasing global prosperity and economic freedom (Berggren and Nilsson, 2015).
Another strand of research suggests that intolerance may be fuelled by the opportunistic behaviour of political actors seeking to advance their own personal interests. Contemporary radical right politicians who seek to gain political advantage from intolerance must also achieve reconciliation with the broadly tolerant context of the prosperous, open, liberal democratic societies in which they operate. Previous studies have observed that radical right parties often utilise tolerance for strategic reasons: to broaden their appeal among voters (Almeida, 2017); to close the 'radical right gender gap' by appealing more equally to men and women (Mayer, 2015); and, in the case of the British National Party's recruitment of Sikhs and fielding of ethnic minority candidates, to 'counter charges of anti-Semitism and racism' (Goodwin, 2010: 178). We agree that association with fascism and Nazism may undermine a radical right organisation's popular legitimacy and that a radical right organisation that aims to recruit male and female supporters would do well to have policies that appeal to both sexes.
We contend that a radical right organisation may express tolerance towards previous out-groups to reduce the level of stigma that society imposes upon its members and supporters. In a generally tolerant, liberal environment, to openly hold male chauvinist, homophobic, biologically racist and Islamophobic views is to adopt a highly stigmatised position. For many individuals, the costs of such stigma may be so great that they will outweigh the benefits of supporting the organisation. Radical right organisations can reduce the burden of stigma by modifying their public position across a number, but not all, of these dimensions: the organisation may publicly pronounce in favour of women's rights, LGBT people and some racial minorities, while maintaining an Islamophobic position, and thereby significantly reduce the costs of stigma. While support is often expressed for women, LGBT people and Jews, radical right organisations do not advocate for all historic out-groups, with Black people and gypsies/travellers being notable omissions. One reason is that such universal tolerance would essentially unravel what it means to be a radical right group. A radical right group, if not intolerant of all out-groups, must be intolerant of at least one out-group. But which out-group? Because more recent political history identifies Islam as salient and perceived as the principal threat by radical right groups, it is targeted as an out-group. 2 The contextual factors that underpin selective tolerance are therefore more complex than mere political appeal and stigma reduction. An out-group must be available upon which intolerance can be targeted, which simultaneously allows the group to reduce stigma, broaden appeal and yet still hold true to what it means to be a radical right group. Crucially, radical right groups express tolerance towards previous out-groups with reference to Islam's alleged intolerance of these groups. It seems rare that radical right groups make statements calling for LGBT rights, gender equality and so on independent of referring to Islam. As Spierings and Zaslove (2015: 143) describe, gender-and LGBT-equality may have become Islamophobic tools so that 'attitudes regarding gender equality and lesbian and gay emancipation . . . might be associated with the nativist, anti-immigration, anti-Islam agenda in a mutually reinforcing manner'. This relationship of mutual reinforcement underpins selective intolerance and provides a powerful tool that can be used to explain why radical right organisations only display tolerance towards certain historic out-groups, and why radical right organisations operating in contexts where there is not a significant concern about stigma do not display selective tolerance but continue to express intolerant views towards a multitude of traditional out-groups.

Modelling selective tolerance
The model to be presented here follows Glaeser's (2005) political economy model of hatred. In Glaeser's core model, pro-redistribution and anti-redistribution politicians compete for votes. If the out-group is generally poor relative to the in-group (e.g. Blacks in the American south in the late 19th and early 20th centuries), anti-redistribution politicians will have an incentive to spread hate about the out-group. Why? Because if relatively poor voters (e.g. working-class Whites in the American south) believe the hate stories, they may vote for an antiredistribution politician who would economically weaken the poor out-group. In this way, some relatively poor in-group voters decide to vote for the anti-redistribution politician although they would have been better off voting for a pro-redistribution politician in the absence of an out-group, or if they believed the out-group to be harmless. Given that they believe a harmful out-group to exist, the gains from weakening them (by reducing their resources) and the harm they can do outweighs the resource gains that would have accrued from voting for a pro-redistribution politician. In this way, the anti-redistribution politician can expand their support base beyond the relatively well-off, by spreading hate stories about a relatively poor out-group. Glaeser argues that the logic is the same if an out-group is perceived as rich such as Jews. Then, pro-redistribution parties can obtain support from relatively wealthy members of the in-group who support taxing and weakening the hated rich out-group. 3 Glaeser acknowledges that the policy dimension and the issues that divide in-group from out-group need not be economic and this is the angle that will be explored here. In particular, the focus here is on two out-groups with views that potentially conflict even more sharply between each other than they do with the in-group. As a result, the policy dimension to be politically contested may specifically regard rights accruing to one of the out-groups. As rights increase (thus helping one of the out-groups), the other out-group, in response, may be made worse off. So, for example, suppose the policy focus is on LGBT rights. An increase in LGBT rights while obviously strengthening the LGBT out-group may also serve to weaken members of a socially conservative out-group such as traditional Muslims who prefer group-based, multicultural rights to trump the granting of individual rights such as gay marriage. In this case, a radical right politician who is relatively socially liberal will have an incentive to spread lies about the conservative out-group. Note that we are analysing the internal competition of a radical right party, so the relatively liberal politician will still be considered far from liberal in terms of the preferences of wider society. They are liberal compared to their fellow radical right members. If the lies are not investigated by a proportion of the radical right membership, a subset of these members who are conservative may vote for a more liberal politician. The meaning of a lie here would be that Muslims wish to inflict damage on society which would include the radical right members. Damage could be that all Muslims are terrorists who wish to impose sharia law on society. Although they would not otherwise wish to see the extension of individual rights, they perceive the Muslim conservative out-group as harmful and will be made less harmful by a society which prioritises individual over group-based rights. So long as the perceived weakening of the Muslim conservative out-group provides benefits that outweigh the cost of a more liberal society, then relatively more conservative radical right members will have an incentive to support relatively liberal radical right politicians. In this way an attempt to align with LGBT causes is driven by the perception that LGBT people are less of a threat compared to the Muslim community. Part of the fear of the Muslim community is a fear of multiculturalism which is more unifying than the fear of LGBT causes.
In terms of the distinction we have been highlighting between traditional radical right groups with widely dispersed intolerance towards a range of out-groups and radical right groups containing a more concentrated intolerance towards a specific out-group, the following model can be viewed in this way. Again, suppose the traditional radical right group is feeling pressure due to social stigma emanating from wider society. They have two goals. They need to retain membership by reducing stigma, but still constitute the identity of a radical right group which requires intolerance of an out-group even if intolerance of all out-groups is no longer possible. A way to satisfy both aims is to attack the traditional conservative out-group (say the Muslim community) for being hostile towards liberal causes (say LGBT rights). In this way, the reduction in stigma may allow retention of relatively liberal radical right group members (although they are unlikely to be considered liberal by wider society) because they feel the pressure of social stigma more keenly than the more conservative members of the radical right group. Wider society while stigmatising intolerance of all groups may be focused more on stigmatising intolerance of groups that hold modern, liberal values than groups that hold more traditional, conservative values. In addition, the group by attacking the Muslim community on the basis of their perceived hostility to liberal rights allows the radical right group to be more cohesive. Traditional, conservative members of the radical right group (who are likely to be less concerned about social stigma) can be persuaded to support the strategy of being selectively liberal (who would not support the strategy otherwise) as it still allows for the intolerance (against Muslims) that is essential to be a radical right group and they view the strategy as hurting the Muslim community. In this sense they are analogous to the poor White voters in Glaeser's application who support anti-redistribution candidates in the presence of a perceived threat from a poor out-group, which they would otherwise not support. The necessity to appeal to wider society also explains why the radical right would choose to align with a liberal social cause, although on social issues they may be closer to the Muslim conservative out-group.
The model follows that of Glaeser with modifications. There are two out-groups, the LGBT and Muslim communities, but we will focus only on the potential harm that can be done by Muslims. Policy towards the LGBT community will provide the focus for political competition. Muslims are harmful with probability h and harmless with probability 1 -h. Harm could be viewed as a belief that the Muslim community wish to engage in terrorism and impose Islamic values upon society. If harmless, no such collective desire exists to do this, in any serious form, within the Muslim community. If harmless, the radical right will suffer no damage but if they are harmful will suffer damage equal to d -(x Px C ) 2 , where d > 0, x P is the radical right policy towards rights and x C is the ideal approach to rights from the perspective of the Muslim community. The squared term equalises positive and negative deviations from x C . Damage is the disutility that would be caused by terrorism and/or the imposition of Islamic values. In this formulation, the damage that can be caused by Muslims towards the radical right is reduced the further that policy is from x C . We assume that d À ðx P À x C Þ 2 ! 0. Signals of potential harm can be received either because Muslims are actually harmful, or as a result of lies spread by radical right politicians and the source of the signal is unclear. We do not consider the scenario in which Muslims are actually harmful, as we wish to address the possibility of increased intolerance towards Muslims being connected to increased tolerance towards the LGBT community in a setting where intolerance is not justified by their actually being harmful. We assume for simplicity that spreading lies is costless for radical right politicians and, therefore, all such politicians will do it. As a result, if a false signal is sent by a radical right politician and received by members of the radical right, members believe that the group is harmful with probability h.
Glaeser considers costs of self-protection from potentially harmful groups. Here, we consider psychological costs which we label d. The idea here is that lack of trust in the Muslim community can be a source of distress. Trust can be established by discovering that the Muslim out-group is, in fact, harmless. The larger is d the larger the distress. Having received a signal, in-group members decide whether to investigate if the signal is true. For that they incur a cost s which is randomly distributed. Whether the search is conducted is determined by the expected gain of undertaking it. If no search is conducted, the expected cost is hðd À ðx High psychological costs caused by d will incentivise search, as will having a low s. The latter might be caused by education, but the presence of d captures the idea that even if someone has a low s they may not search if d is low or even negative. These individuals may enjoy feelings of distrust towards the out-group. 4 For others though, a failure to search may simply be that it is too costly. Search is also likely if h is small. The share of the radical right membership that will search is therefore given by Hðdð1 À hÞÞ. So, the share of the group that will not investigate the truth is 1 À Hðdð1 À hÞÞ.
Glaeser considered income (y) as the policy issue such that if a tax rate is proposed, an individual with income y will pay sðy ÀŷÞ in tax where s is the tax rate andŷ is average income. If one politician argues for higher s, the relatively wealthy will pay more tax under this proposal and relatively poor individuals will receive a greater subsidy. Here, instead the fight will be over rights. Ignoring hatred a radical right member i will support the radical right politician proposing conservative group rights so long as ðx i À x con Þ 2 < ðx i À x lib Þ 2 , where x con and x lib are the positions of the relatively conservative and liberal radical right politicians, respectively. This simply says that the radical right group member, in the absence of hatred, would support the politician offering a position on rights that is closest to their own. The indifferent group memberx would be such that If we assume that x lib < x con , the proportion of support within the group for the liberal politician is FðxÞ.
The existence of hatred may change the direction of support for certain types of radical right members. If lies have been spread and not investigated by some group members, then there is a belief that Muslims could do harm to society. However, policy proposals made by the competing politicians may reduce this harm and crucially the politician who proposes the policy that will weaken Muslims the most will have a competitive advantage. To see this, suppose reasonably enough that the radical right is on average a traditionally socially conservative group but that Muslims are, on average, even more socially conservative and less keen on the extension of individual rights. A policy that extends individual rights would weaken Muslims in the event that they are harmful. 5 Someone who chooses not to investigate the hate message (the haters) will support the politician arguing for relatively liberal policies if We can see that for some group members, who would in the absence of hatred have chosen a socially conservative policy, will in the presence of hatred choose a more liberal approach to rights. The reason is that the benefit of the relatively greater damage that this policy inflicts upon the socially conservative Muslim out-group outweighs the loss in terms of an approach to rights that is further from their ideal point. In this way, a more liberal approach which recommends extending individual rights to the LGBT community could be combined with hate messages aimed at Muslims. If the hate messages are believed, the relatively liberal radical right politician within the group will no longer just collect support from other relatively liberal group members. In addition, they can pick up additional support from relatively socially conservative group members who see a liberal extension of rights as a way of weakening the Muslim out-group that they fear. Among those that hate the indifferent group memberx h has preferences that are more conservative thanx (in equation (2)) as shown bŷ We can see thatx h >x given that 2x C À x con À x lib > 0, which follows from the assumption that x C is further from x lib than x con . The additional support to be gained by spreading hate is ð1 À Hðdð1 À hÞÞÞðFðx h Þ À FðxÞÞ which is positive because Fðx h Þ À FðxÞ > 0.
If the out-group were to become more conservative (x C increases), x h increases and the incentive to spread hate by the radical right liberal politician increases. This reflects the fact that in the mind of the haters, the liberal politician can inflict even greater damage on the Muslim out-group as the distance between them and Muslims becomes even greater. If h increases, hate increases. The reason is that if the probability of the out-group actually being harmful increases, there is less incentive to search to see if the hate messages are true. If d increases hate decreases, this reflects an increasing psychological discomfort from feeling distrust towards Muslims. This could be related to closer integration of groups. Finally, if s increases across the in-group, hate increases because search for the truth is costlier. This could be related to education and/or levels of integration.
We emphasise that social stigma plays a key role here, though it is hidden. The false messages are aimed at Muslims only and this would happen when, for example, homophobia or sexism is highly socially stigmatised. The reason is that the relatively more liberal members of the radical right would be feeling the pressure of social stigma. This pressure could be felt, for example, in job market concerns, that if viewed as homophobic they would be discriminated against (Becker, 1971). In addition, recruitment will be easier for the radical right if highly stigmatised views are suppressed. Were homophobia and sexism not stigmatised, we would also expect false messages to be sent about gender equality and the LGBT community. To that extent, false messages from within the different spectrums of the radical right group would tend to counteract and perhaps neutralise each other. The focus on one group (in this case Muslims) stems from an environment where less social stigma is attached to attacking this out-group than other traditional out-groups. Formally, we could think of this being captured by differences in the receipt of hate messages with reference to equation (1). For example, if a hate message is received about the LGBT community in a society where homophobia is highly socially stigmatised, d may be higher and h and s lower than for Muslims, thus inducing much greater search for the truth than for false messages concerning Muslims. We now turn to case studies to provide evidence for this mechanism in the context of intolerance towards Muslims.

Case studies
If our analysis is correct, then we would not expect to see a radical right group couch its Islamophobia in liberal values if it is operating in a society in which little or no stigma attaches to intolerant views. However, we would expect to observe a radical right organisation expressing selective tolerance when: (1) expressing tolerance may be a means of reducing the stigma attached to their other intolerant views and (2) expressing tolerance will impose costs on Muslims. We would also predict that as a radical right organisation begins to champion the causes of former out-groups, its anti-Islam rhetoric will increase. Indeed, its championing of former out-groups is couched in relation to its negative depiction of Islam and not in isolation from it.
We will examine these hypotheses via analysis of case studies from contemporary Polish and UK politics. Poland historically has a patriarchal culture and there remains a strong commitment to traditional gender roles. Women's emancipation is associated with state intervention during the socialist era and therefore became discredited in the post-socialist period (Lehoczky, 2005;Saxonberg and Szelewa, 2007). The cost of holding homophobic views may be lowered by the lack of legislation that allows prosecution for homophobic hate speech; although the Polish Penal Code does include some provisions related to hate speech, neither sexual orientation nor gender identity are protected characteristics (Article 19, 2018). In addition, Poland's Constitution does not explicitly contain provisions that prohibit advocating hatred 'that constitutes incitement to discrimination, hostility, or violence' (Article 19, 2018: 17). Same-sex marriage is prohibited and there are widespread levels of homophobia, with 42.5% of Polish respondents to the World Values Survey (2015) agreeing that homosexuality can never be justified. Poland is also the most anti-Semitic country in Eastern Europe (Anti-Defamation League, 2015). Finally, there are extensive levels of Islamophobia, with 67% of respondents reporting antipathy to Arab people (CBOS Public Opinion Research Centre, 2016). Given the evidence of intolerance towards all out-groups, we would not expect a political organisation to broaden its appeal or reduce stigma by championing the causes of any minority. Intolerant actors within Poland will not be forced to make a trade-off; there is no need for a radical right organisation to express tolerance towards LGBT people or Jews in order to express intolerance towards Islam. However, the UK population is increasingly tolerant towards the radical right's historic out-groups. A 2015 poll from the Pew Research Centre found that 7% of UK respondents had an unfavourable view of Jews. Workplace discrimination against women in the United Kingdom is outlawed and, although a gender pay gap still exists, it has narrowed in recent years (OECD, 2017). There is also increased tolerance for homosexuality: although 74% of respondents in 1987 agreed that same-sex relationships were 'always' or 'mostly' wrong, that number dropped to 19% by 2016(British Social Attitudes, 2016. The cost of engaging in homophobic practices or expressing homophobic views has increased since the UK Parliament outlawed discrimination on the grounds of sexual orientation in 2007. The following year it became illegal to express homophobic hatred. Clements and Field (2014: 543) note that increasingly liberal attitudes towards homosexuality 'aligns with growing popular support in Britain for equality and diversity in all forms, but particularly in relation to gender, race, and religion'. The social research agency NatCen has found a significant and steady reduction in the percentage of respondents who say that they 'would "mind a lot" or "mind a little" if a close relative married someone who was Black or Asian' (Kelley et al., 2017: 10). NatCen has also observed that views on gender roles -measured by asking whether respondents agreed that it is a man's job to earn money and woman's job to look after the home and family -have become more progressive over time and particularly since 2008 (Phillips et al., 2018). Although there has been a troubling rise of anti-Semitic incidents reported throughout Europe, recent YouGov (2017) polling within the United Kingdom found that 81% of respondents agreed it was fair to administer a lifetime game ban to football fans who made Nazi gestures which suggests that expressions of anti-Semitism are stigmatised.
However, Clements and Field (2014) also observe that an exception to the broader trend of tolerance is Islamophobia, which has grown as intolerance towards other historic out-groups has decreased. Islamophobia in the United Kingdom increased after the 9/11 and 7/7 terrorist attacks, and a YouGov (2016) poll found that over half of respondents agreed there is a fundamental clash between the values of Islam and British society. Therefore, a radical right organisation that expresses tolerance towards previous out-groups while also advocating for Islamophobia will plausibly attain broader appeal and encounter less stigma than a consistently intolerant party. This is particularly so given the increased political salience of Islam: in 2000 survey evidence showed that only 17% of UK respondents were unwilling to accept any people from Muslim countries who wished to work in the European Union (Field, 2007: 452); however, a 2017 survey revealed that 47% of UK respondents agreed that '[a]ll further migration from mostly Muslim countries should be stopped' (Goodwin et al., 2017).

Poland and broad intolerance
Since 2015, Poland has been governed by the Prawo i Sprawiedliwo s c, or Law and Justice (PiS), a populist right party that has called to preserve the Polish nation through traditional Catholic values (Gwiazda, 2016). After winning an unprecedented majority, PiS has sought to weaken core democratic institutions such as the media and Constitutional Court (Bustikova, 2018). We hypothesise that if PiS's local operating context is not broadly tolerant, the party will not display selective tolerance. Instead, we would expect intolerant actors to express intolerance towards all out-groups, including Islam, and we would not expect criticism of Islam to be couched in support of liberal values.
Public statements of PiS representatives support this hypothesis. PiS's foreign minister expressed biological racism when criticising the previous government for having promoted 'a new mixture of cultures and races' (Easton, 2016). The European Jewish Council has warned that anti-Semitism in Poland has been normalised and that anti-Semitic incidents have increased since PiS closed communications with official representatives of the Jewish community (Halon, 2017). In the lead-up to the 2019 election, PiS's Chair, Jaroslaw Kaczynski, maintained that LGBT rights threatened Polish values and families (Easton, 2019). The party also release a campaign advertisement that featured a PiS-branded umbrella shielding a family from rainbowcoloured rain (Noak, 2019).
In addition, PiS displays Islamophobia. Consistent with its calls to uphold traditional Catholic values, PiS have attacked Islam and claimed it may harm Poland's religious identity, with Kaczynski, stating: [Muslims] do not follow our laws and customs and then impose their sensitivity and requirements in public space in various areas of life. And in a very aggressive and violent way . . . Let's look at Sweden, the 54 zones where sharia is in force. Fears of hanging the Swedish flag at schools because there is a cross. Swedish students are not allowed to wear short dresses because they do not like it . . . What's going on in Italy? Stained churches, treated as toilets. (Niezalezna, 2015) Rather than criticising Islam for oppressing women, LGBT people and Jews, PiS attack Islam primarily on the grounds that it may pose a security threat, harm Poland's Catholic identity and threaten the femininity of women rather than their rights.
Radical right organisations in Poland have also attracted tens of thousands of people to their street-based protests which promote a homophobic, anti-Semitic and Islamophobic ideology. The All-Polish Youth, an ultra-Catholic nationalist organisation with an anti-Semitic history, is a co-organiser of the Polish Independence Day march. In 2017, the march attracted a near-record 60,000 participants and included the chanting of anti-Semitic, anti-Muslim and anti-gay slogans (Pikulicka-Wilczewska, 2017;Taylor, 2017). Because the All-Polish Youth is not operating in a broadly tolerant environment, it will not have to couch its prejudice in liberal values to attract support and can express opposition to historic out-groups without fear of jeopardising its base.

The United Kingdom and selective tolerance
The UKIP was founded in 1993 with the intention of contesting elections to secure the United Kingdom's withdrawal from the European Union (Usherwood, 2008). However, single-issue Euroscepticism was not enough to drive voters towards the party. To increase its share, UKIP developed 'a more concerted voteseeking strategy', with its leader, Nigel Farage, seeking to merge Europe and immigration in the minds of voters (Ford and Goodwin, 2014: 282;Tournier-Sol, 2015).
As the party has moved to a platform more closely associated with the populist, radical right it has been dogged by increasing allegations of racism and xenophobia. UKIP has been more successful in attracting voters who are less likely to view non-White citizens as British and more likely to view Islam as a serious danger to Western civilisation than any other major UK party (Ford et al., 2012;Goodhart, 2014). In addition, there have been numerous high-profile racist, sexist and homophobic gaffes associated with the party (Hatakka et al., 2017;Kaufmann, 2014). Ford et al. (2012) identified the dilemma that may precede the adoption of selective tolerance when they recognised that [I]f UKIP continues to put strong emphasis on divisive issues such as immigration and Islam then it risks alienating strategic supporters who are primarily motivated to defect from the Conservatives by their Euroscepticism. In the future, much will depend upon how the party manages this difficult trade-off. (p. 228) We take the view that UKIP attempted to navigate the trade-off via the strategic considerations that underpin selective tolerance and contend that the party has expressed tolerance to reduce the stigma attached to their other intolerant views and to impose costs on Muslims.
A YouGov (2015) poll found that 55% of UKIP voters would be embarrassed to tell their family and friends about their voting intentions. There is stigma attached to open endorsement of UKIP which may have been fostered by its anti-immigrant rhetoric and the sexist and homophobic comments by its candidates and councillors. Expressing tolerance for historic out-groups may be a way to minimise the stigma associated with UKIP support.
We would expect UKIP to justify its intolerance towards Muslims in accordance with liberal values. Its 2017 General Election manifesto states that 'UKIP will ban wearing of the niqab and the burqa in public places . . . We will not accept these de-humanising symbols of segregation and oppression . . . ' and claims that 'anyone who believes in women's rights, should be outraged by the appalling practices occurring on a daily basis in minority communities across Britain' (UKIP, 2017: 36, 37). The manifesto quotes UKIP's LGBT chair, who claims that '[t]hose of us who identify as lesbian, gay, bisexual and transgender may have first-hand experience of how misogynistic and homophobic attitudes are tolerated in the name of "respecting cultural differences" ' (UKIP, 2017: 37). In addition, the manifesto requests Ofsted, the UK regulator of schools and colleges, to conduct inspections of schools where staff or governors are believed to espouse 'anti-Semitic . . . views' as part of UKIP's policy to end Islamist extremism in schools (UKIP, 2017: 38).
As UKIP's defence of women, LGBT people and Jews has increased, so too has its Islamophobia. Its 2001 General Election manifesto contained no relevant references to the former out-groups or to Muslims but reflected its essentially Eurosceptic nature with 228 references to the European Union. By contrast, its 2017 manifesto contained 16 references to Islam and Muslims, 29 references to women, 12 references to Female Genital Mutilation, two references to LGBT people and one reference to anti-Semitism, while its references to the European Union dropped to 153. It seems that when Islam was not politically salient, neither were women, gays or Jews.
In the aftermath of the 2016 UK referendum vote to leave the European Union and their dismal performance in the 2017 General Election, UKIP intensified their hostile focus on Islam to such an extent that in December 2018, Nigel Farage resigned as a party member in protest at the appointment of former EDL leader Tommy Robinson as a party adviser. Statements made by the party's leadership reveal that its anti-Islam stance is premised on more than a rejection of multiculturalism and instead suggests that the party is seeking to inflame antipathy towards Muslims through the spreading of hate; one of Farage's successors, Gerard Batten, described Islam as a 'death cult' in an interview with Sky News (2018) and stated that, '[t]he ideology of the religion does actually traditionally throughout its history make sex slaves legitimate. Their prophet had sex slaves . . . It's in the Hadith and the Koran. It's a tradition that that's okay to do that'.
The EDL, a street-based protest organisation founded in 2009 to oppose the spread of radical Islam in the United Kingdom, is also selectively tolerant. The EDL has a female division for its members and from the outset emphasised that it is a multicultural organisation that includes Sikh, Jewish and LGBT people (Allen, 2011). From July 2010 to November 2011, the EDL held almost monthly protests that attracted between 500 and 2000 participants. At many of these demonstrations participants carried rainbow or Israeli flags and encouraged the attendance of Sikhs, while engaging in anti-Muslim chanting. We argue that the EDL's promotion of women, LGBT people, Sikhs and Jews is the result of the strategic considerations of selective tolerance. Meadowcroft and Morrow (2017) have observed that EDL participation is a stigmatised activity with much of the encountered stigma being derived from the public perception that the group is racist. In a bid to challenge this perception, many EDL members were at pains to portray themselves as anti-racist and tolerant of minority groups. This would often be done by members claiming to have family and friends of different races or, as one young member put it, 'I'm not racist. I don't have a problem with Sikhs or Hindus or Jews or gays and lesbians' (Meadowcroft and Morrow, 2017: 384).
Original ethnographic data reveals that EDL leaders additionally promoted the organisation's inclusion of historic out-groups to provide a rebuttal to people who might accuse participants of racism. At one meet and greet session for new members, EDL leaders told the audience that the EDL has divisions for Jews, Sikhs and LGBT people and that 'if anyone says that the EDL are racist you should the let other person know that there is a division for all different religions'.
As our model anticipates, the EDL's tolerance of women, LGBT people and Jews enables the organisation to simultaneously impose costs on Islam by singling out the religion for its alleged illiberalness and intolerance. Speakers at EDL demonstrations warned the audience that if Islamists 'have their way gays and Jews will be stoned and so will women'. In addition, the EDL's mission statement claims that 'other expressions of Islam-inspired intolerance . . . include: denigration and oppression of women, organised sexual abuse of children, female genital mutilation, so-called honour killings, homophobia, racism, anti-Semitism' (EDL, 2016). Like UKIP, the EDL (2011) specifically targets adherents of the Islamic faith via inflamatory rhetoric, with the organisation's website stating, The importance of identifying Islam as a major contributing factor is demonstrated by the arrests for child-grooming that have recently swept across the country . . . many Muslim men see little wrong with applying the example of the prophet (sex with young children) to those who they regard as 'dirty kuffar'.
The strategic nature of the EDL's selective tolerance is revealed in its reluctance to advocate for all historic out-groups. The organisation does not have a designated division for black members, despite having divisions for ethnic minorities, such as Sikhs, that are far less significant in terms of the UK population. It seems plausible that the EDL does not have a division for Black members because advocating for Black people would not impose costs on Muslims.
The strategic considerations that underpin the EDL's selective tolerance are not lost on the organisation's supporters. This is illuminated by an exchange on the anti-Islam 'Gates of Vienna' blog after its posting of the EDL's mission statement. One poster contended, 'And here's the problem with the "Counter-Jihad" movement: it's all built on the basis of being opposed to something, rather than being for something. Geert Wilders bases his opposition to Islam around abstract ideas like freedom, but really, what does that mean? Does that mean we oppose Islam so we can have our gay marriages, pot smoking, and abortionon-demand?' In response, another poster observed: '. . . You make some pertinent points. My argument is that "purity" will follow change, and that in the meantime any action that brings whites together in common cause against the establishment is good -even if it is "tainted"'. 6 This exchange suggests that selective tolerance can be used as a tool by a radical right organisation to maintain and expand its supporter base. However, if tolerance is extended too broadly, this may undermine the radical right organisation's raison d'etre and alienate supporters.

Conclusion
This article has set out an explanation for the selective tolerance exhibited by contemporary radical right organisations. Where intolerant views are stigmatised, radical right organisations have sought to reduce this stigma by expressing tolerance towards some historic out-groups. Where intolerant views do not encounter stigma, there is little evidence of selective tolerance. Where radical right organisations exhibit tolerance, it cannot be expressed universally. Such broad expression of tolerance would essentially undermine the rationale for being in a radical right group. So, at least one group must be subject to intolerance. In addition to reducing stigma, it is advantageous if the sort of tolerance that is expressed by the intolerant imposes costs upon the targeted out-group. Where such selective tolerance is witnessed, in recent times Muslims are the chosen out-group and support for the rights of women, LGBT people and the Jewish community is judged to both detoxify the radical right group and impose costs on Muslims.
One can speculate what might happen if the current climate were to change in Western Europe and Muslims became less salient as the obvious out-group to target. Would there be a return to broad intolerance? Or would a tolerance for traditional Islam be fostered as a tactic to target alternative out-groups such as Jews or the LGBT community? Indeed, one might argue as Cohen (2007) has, that it is this sort of selective tolerance that is characteristic of radical left rather than radical right groups. In these cases, tolerance for groups that one would not normally associate with radical left causes has arguably occurred because these groups are anti-US and/or anti-Israel.
Likewise, in the approach taken here we argue that a move towards individual and away from group rights weakens Muslims. Some would argue the opposite may be true as Muslims feel victimised by the action of the ingroup; they actually pose a greater threat. However, among those that truly hate the same argument applies. Haters would believe that giving socially conservative Muslims what they want and strengthening group-based rights would increase the threat from Muslims. 6. Comments on Gates of Vienna blog, 15 January 2011.