Angry White Men: Individual and Contextual Predictors of Support for the British National Party



The British National party (BNP) is the most successful extreme right party in Britain's electoral history and is the fastest growing political party in twenty-first century Britain. This article presents the first ever individual-level analysis of BNP supporters, utilising a survey data set uniquely compiled for this purpose. We find that support for the BNP is concentrated among older, less educated working-class men living in the declining industrial towns of the North and Midlands regions. This pattern of support is quite distinct from that which underpinned the last electorally relevant extreme right party in Britain – the National Front (NF) – whose base was young working-class men in Greater London and the West Midlands. Extreme right voters in contemporary Britain express exceptionally high levels of anxiety about immigration and disaffection with the mainstream political parties. Multi-level analysis of BNP support shows that the party prospers in areas with low education levels and large Muslim minority populations of Pakistani or African origin. The BNP has succeeded in mobilising a clearly defined support base: middle-aged working-class white men anxious about immigration, threatened by local Muslim communities and hostile to the existing political establishment. We conclude by noting that all the factors underpinning the BNP's emergence – high immigration levels, rising perceptions of identity conflict and the declining strength of the cultural and institutional ties binding voters to the main parties – are likely to persist in the coming years. The BNP therefore looks likely to consolidate itself as a persistent feature of the British political landscape.

The British National party (BNP) is the most successful extreme right party (ERP) in British electoral history and is currently the fastest growing political party in Britain. Founded in 1982 by a former chairman of the party's predecessor, the National Front (NF), BNP support has been on a rising trend ever since, with a particularly sharp increase in the last general election when the party quadrupled its vote (see Table 1). Unlike earlier ERPs in Britain which achieved only ephemeral success and failed to gain more than a handful of elected representatives, the BNP has now established a sizeable legislative base; the party currently has over 50 incumbent local councillors, two Members of the European Parliament and one seat on the Greater London Assembly (GLA).1 The BNP's rise has sparked considerable political and academic debate. Some, such as Labour MP Jon Cruddas, suggest that the extreme right has become a ‘home for many disgruntled former Labour voters’, and that the BNP is ‘beginning to establish itself as a rival to Labour in many of our traditional heartlands’ (Kite, 2006). Other Labour MPs interpret the BNP's rise as ‘a result of local political failure’, suggesting that the party performs well on ‘estates that have been ignored for decades’ and among ‘white skilled working-class voters who feel politicians live on a different planet’ (Blears, 2008, p. 42). Pointing towards severe economic recession and public concern over immigration others compare current BNP voting to an earlier wave of extreme right voting for the NF when similar conditions existed in the 1970s (e.g. Travers, 2009).

Table 1. Support for the Extreme Right in General Elections, 1970–2005
Election yearVotes receivedConstituencies contestedAverage in seats contested (%)Highest vote (%)Deposits retained
  1. Notes:aAn unofficial candidate from the NF Flag Group contested the constituency of Bristol East.

  2. b Officially the BNP leadership abstained from the 1987 general election though two unofficial candidates still stood.

  3. c In addition, a rival extreme right party, the National Democrats (ND), fielded 21 candidates who combined received 10,829 votes and averaged 1.2 per cent (the highest vote was 11 per cent).

  4. Sources:Eatwell (2000, p. 173); Taylor (1982); Yonwin (2004).

1974 (Feb)76,865543.27.80
1974 (Oct)113,843903.19.40
1987    0

The social and attitudinal drivers of BNP support remain, however, under-researched. While there is a burgeoning pan-European literature on contemporary ERPs (for a recent overview see Mudde, 2007), we still know relatively little about who votes for the extreme right in Britain and why. In this article we analyse the first ever representative data set of individual BNP supporters. Our goals are threefold: first, to examine the social profile and political attitudes of individual BNP supporters; second, to compare these supporters with those of the only previous British ERP of electoral relevance, the National Front; and third, to test whether contextual conditions, in particular economic deprivation and ethnic diversity, are associated with higher BNP support after controlling for individual characteristics.

We find that BNP supporters have a clear social profile: they are predominantly middle-aged working-class men with few educational qualifications. Although they are concentrated in working-class jobs, BNP voters are not particularly poor – they are no more likely to live in government-provided housing, and are only slightly more likely to be unemployed. BNP voters also have clear political preferences: they are very worried about immigration and profoundly hostile to the political establishment. Moreover, the modern-day BNP is quite distinct from the 1970s NF: its support is older, more Northern and more disaffected with mainstream politics. BNP supporters also concentrate in particular social contexts: the party's support levels are higher in constituencies with low education levels and large Muslim minorities of Pakistani or African origin. While we cannot be sure that the BNP is drawing support from Labour, the social and geographic contexts where BNP voters concentrate also show elevated support for Labour. We conclude that the BNP has succeeded in establishing itself by mobilising a distinct population of white working-class voters threatened by immigration and ethnic change and profoundly dissatisfied with the political status quo. Furthermore, the social and political trends that underpin the emergence of the BNP – high immigration, the perception of conflict between white and Muslim communities and the erosion of institutions and traditions binding social groups to the Labour party – are likely to continue. We therefore expect the BNP, or a similar successor, to become established as a fixture in mainstream British politics and a competitor to the Labour party in areas where the latter has faced little competition before.

The Resurgence of Extreme Right Politics in Britain

During its early years the BNP languished in the electoral ghetto and was more noteworthy for issues of public order than electoral relevance. The early BNP never fully committed itself to electoral politics, adhering instead to a strategy of ‘very limited involvement’ in elections which led activists to describe it as the ‘no-to-elections party’.2 In the 1983 general election the BNP fielded candidates in over 50 constituencies yet delivered election addresses in only five, the party officially abstained from the 1987 general election and in 1992 it stood only 13 candidates.3 During its first two decades the BNP acquired just one local councillor. Rather than contest costly elections, the BNP instead adhered to the British extreme right's traditional tactic of ‘march and grow’, using confrontational street-based rallies to raise its profile and attract new recruits.

Since the election of Nick Griffin as party chairman in 1999, the BNP has focused more strongly on building electoral support, with impressive results. In the decade since Griffin's election, the BNP has become the most electorally successful ERP in British history. In the European elections of June 2009 the BNP secured two seats in the European Parliament and almost 1 million votes, a nine-fold increase on the party's result in 1999 and the largest total for an ERP in any British election. This result was the culmination of a long upward trend. In 2008 the party won its first representation outside local government by polling over 5 per cent of the list vote in elections to the Greater London Assembly (GLA). In the general election of 2005 the BNP more than tripled its number of candidates to 119 while its total vote quadrupled to over 192,000 (Table 1). The BNP has also secured more than 200,000 votes in each round of local elections since the 2005 general election, and currently holds more than 50 council seats.

How might we account for this rapid rise in BNP support? Different hypotheses can be derived from three theoretical frameworks used to explain vote choice in contemporary Britain: valence voting (Clarke et al., 2004; Stokes, 1963); spatial or ideological voting (Budge, 1994; Downs, 1957); and expressive or identity voting (Green et al., 2004; Heath et al., 1985; 2001). Valence voting occurs when voters (and parties) agree about what they want the system to deliver and political choice is a decision about which party is best equipped to deliver these results. Such a framework would attribute the emergence of a new party to a persistent failure by the existing parties to convince voters that they can deliver the consensus outcome in particular policy areas. The new party should be able to offer more credible management in these areas, and those who defect to it first should be those who are most concerned about the policies where failure has occurred.

The ideological or spatial voting framework argues that fundamental disagreements over values and priorities do exist among voters, and that in political competition the parties position themselves with respect to these ideological disagreements in an effort to maximise their vote. The classic Downsian prediction is that when competition occurs on one ideological dimension the major parties will maximise their vote by moving away from the extremes towards overlapping positions in the centre ground, where the most votes are. The emergence of a new party generally accepted to be ideologically extreme could have two explanations. The new party may successfully be winning over the more extreme votes at one of the poles of a unidimensional distribution, which may be dissatisfied with the centrist options available (see Kitschelt, 1995). Alternatively, the appearance of ideologically distinctive new parties may suggest the emergence of new conflicts that cross-cut the existing value divide and are therefore not coherently represented by the existing parties. The new party succeeds by being the only party advancing the new ideological agenda.

Expressive or identity voting does not view the choice of parties as a self-interested managerial or ideological calculation. Voters choose a party because of who they are rather than what they want. Political loyalties are rooted in social identities, and shaped by the contexts in which individuals grow up. In Britain, the classic example of this is class voting – voters from particular economic backgrounds retaining throughout life a tendency towards Labour or the Conservatives that cannot be accounted for by their views of these parties' competence or by their ideological priorities. This account would suggest that the emergence of a new party will be rooted in changes to social structure, which either mobilise previously apathetic social groups or lead politically mobilised groups to switch their loyalties. The new party should have a distinctive social profile, with support concentrated in certain core groups and absent from other peripheral groups.

The ‘modernisation’ strategy pursued by the BNP since Griffin's 1999 election as chairman has incorporated elements of all three approaches: a valence politics commitment to responsive community activism, the development of a more moderate and popular ideological appeal which adopts themes from European radical right parties such as the French Front National (FN) and a focus on winning over particular social groups, notably disaffected white working-class voters, supporters of the Labour party in de-industrialising districts and citizens living in or near ethnically diverse areas.

Griffin's BNP has looked towards the French FN and the Liberal Democrats in Britain and developed its own brand of pavement politics, encouraging branches to target local grievances in an articulate manner and urging activists to engage in face-to-face contact with residents (Goodwin, forthcoming [a]). Commencing in London's East End in the early 1990s, BNP ‘modernisers’ refer to the grass-roots tactic as follows: ‘We must stop talking just about what we like to talk about and start talking about the things that local people are crying out to hear’.4 Implicit in the emphasis on intensive, locally orientated campaigns is the party's broader goal of cultivating an image as a legitimate and effective political organisation. Indicative of the BNP's renewed emphasis on community-based politics is the finding in one study that many residents in three northern towns experienced more face-to-face contact with BNP activists than their ‘mainstream’ party counterparts (JRCT, 2004). To promote an image of competence, the party has professionalised its campaigns, urging activists to become better dressed and jettison extremist rhetoric. For example, aside from holding more regular training seminars the party now distributes a Language and Concepts Discipline Manual, advising activists to avoid terms such as ‘fascist’ and ‘racialist’ in favour of ‘right-wing populist’ or ‘ethnonationalist’.5

Ideologically, Griffin's BNP has modified its exoteric appeals, moderating its discourse on race and immigration, and adopting more voter-friendly themes such as ‘Democracy’, ‘Identity’, ‘Freedom’ and ‘Security’ (BNP, 2005). Under Griffin, the party now downplays the crude anti-black racism and conspiratorial anti-Semitism that characterised its predecessors in favour of ‘differentialist racism’ based on the ethno-pluralist doctrine, stating for example: ‘The BNP does not claim that any one race is superior to any other, simply that they are different ... To protect and preserve the racial and cultural integrity of the British people ... the party believes in separation’.6 Alongside this new discourse the ‘modernised’ BNP has abandoned calls to overhaul the democratic system in favour of anti-establishment (but not overtly anti-democratic) populism, resulting in an ideology characteristic of the electorally more successful European ‘new’ radical right (e.g. Rydgren, 2005). The BNP has sought to shed its single-issue status by developing a full manifesto of policies with a clear focus on ethnic nationalism and social conservatism. The appeal of such an ideology has increased in recent years, as issues such as immigration, Islam and law and order have risen up the political agenda (e.g. McLaren and Johnson, 2007; also Eatwell and Goodwin, forthcoming). Echoing findings produced about the NF in the 1960s and 1970s, recent research has shown that the electoral potential of this ‘modernised’ BNP ideology stretches far beyond the party's current support base (Ford, forthcoming).

The BNP has also focused its efforts on particular geographical and social contexts. The party has moved out from the traditional bastions of ERP support such as London's inner East End towards de-industrialising (white) working-class districts in northern England and outer London areas such as Barking and Dagenham where BNP strategists have clearly taken aim at disillusioned ex-Labour supporters. From the late 1990s onwards, the BNP has directed much of its scarce resources towards Labour heartlands where, as a result of a long history of Labour incumbency, party competition is either weak or absent (Wilks-Heeg, 2009). For example, in 1998 the BNP developed new leaflets designed specifically for areas that voted Labour and which also had high levels of welfare dependency, such as parts of the North West and South Yorkshire, and which claimed Labour had ‘abandoned ordinary working people’ and ‘are more interested in “gays”, ethnics and their rich money-bags backers’.7 The BNP's attempt to invade Labour heartlands is aptly reflected in its campaign slogan: ‘We are the Labour Party your grandfathers voted for’ (Collins, 2006), and also in the overall pattern of results: 52 of the 58 council seats won by the BNP since 2005 have come at the expense of Labour incumbents. The party has also targeted areas with large ethnic minority populations, and following the events of 11 September 2001 particularly areas that have, or are close to, large Muslim communities. BNP campaigns emphasise the alleged threat from Islam, presenting Muslims as being incompatible with ‘British’ values and representing a threat to the jobs, culture and demographic dominance of ‘native’ white Britons. As the party itself notes, in recent years its campaigns have focused on ‘the growth of fundamentality – militant Islam in the UK and its ever-increasing threat to Western civilization and our implicit values’.8

Previous Research on British ERPs

ERPs in Europe have been the subject of numerous academic studies, which have revealed distinct social and attitudinal predictors of their support (Arzheimer, 2009, p. 259). Elsewhere in Europe the extreme right electorate tends to be male, with low levels of formal education, generally in manual occupations or the petty bourgeoisie. ERP voters on the continent have also been shown to have a distinct attitudinal profile: they have higher levels of hostility towards non-European immigrants, and are more likely to regard immigration as a source of economic competition and a threat to the national culture. They overwhelmingly prioritise immigration as a pressing political issue, and tend to be very negative about existing immigration policies and about the leadership and policies of the mainstream political parties.

Research into the only other significant ERP in modern British politics – the National Front of the 1970s – has found the NF's support base to be broadly consistent with this wider picture. Aggregate-level research on the 1977 Greater London Council elections, where the NF polled around 120,000 votes, found the strongest predictor of support to be the percentage of manual workers and interpreted this finding as being consistent with the ‘working-class authoritarianism’ thesis (Whiteley, 1979).9 Findings also indicate that the NF benefited from distinct political traditions, with its support concentrated in old East End and transplanted East End communities where economic insecurities arising from the casual labour system of the old docklands rendered voters particularly receptive to anti-immigrant campaigns (Husbands, 1983).10 Higher than average levels of NF voting in areas such as Bradford, Coventry, Leicester, London's East End and Wolverhampton – all of which had high concentrations of early black and Asian immigrants – suggested that the NF benefited from a white ‘backlash’ against rising ethnic diversity (Taylor, 1982). Drawing on these findings some speculated that it was primarily the economically marginal and culturally threatened white worker in working-class districts who was particularly attracted to the earlier NF (Kitschelt, 1995, p. 255).

However, while providing useful insight this body of research is weakened by its reliance on ecological evidence: we learn something about where NF supporters live, but not about who they are or what they think. The task of gathering individual-level data on fringe ERPs is problematic as standard random sample surveys produce too few supporters to permit meaningful analysis. The main attempt to meet this challenge in the 1970s was the study by Martin Harrop et al. (1980), which aggregated 22 commercial polls to produce a sample of 270 NF supporters polled between 1977 and 1978.11 The study suggested a distinctive social and geographical profile to NF support, which they found to be concentrated among young, employed skilled working-class males in Greater London and the West Midlands. NF support was particularly high among skilled rather than semi-/unskilled workers, a finding suggesting NF support was highest ‘in the economic locations where competition between indigenous and immigrant populations for scarce and valuable resources is felt most acutely’ (Harrop et al., 1980, p. 276).12 This study also found NF supporters to hold strongly negative attitudes towards the mainstream political parties, although many expressed less negative views of the Conservatives under Margaret Thatcher.

Studies of the BNP

The BNP's rise has spurred a new series of academic studies, most of which have examined the party's historical and ideological development (Copsey, 2008; Goodwin, forthcoming [c]; though see John and Margetts, 2009). These aggregate analyses suggest that much like its predecessor the BNP draws support principally from areas that are more economically deprived and that have large minority populations. The proportions of skilled manual workers and individuals with no formal qualifications in an area correlate positively with BNP support in local elections, while the proportion of young voters correlates negatively with BNP voting (Borisyuk et al., 2007; John et al., 2006; JRRT, 2005). Unlike the old NF, however, the presence of significant Muslim populations is also identified as an important local factor mobilising BNP support. For example, while Peter John et al. (2006, pp. 16–7) find no relationship between the number of asylum seekers and levels of BNP support, at council level they find a positive correlation between support and the presence of Pakistani (0.380) and Bangladeshi (0.346) ethnic groups, a relationship that does not hold for the presence of Indian Asian groups.13 Benjamin Bowyer's (2008) multi-level analysis of BNP support in the 2002–3 local elections suggested that the BNP was performing most strongly in what might be described as ‘threatened white enclaves’, that is, economically deprived homogeneously white wards within ethnically diverse local authorities with large Muslim populations.14

The existing research suffers from several shortcomings which we aim to address. Firstly, previous studies are driven by either aggregate analysis or qualitative interviews with small non-random samples of BNP activists (Goodwin, forthcoming [b]) or voters (Rhodes, 2006). Prior to our study there has been no statistical examination of a representative sample of individual BNP supporters. Secondly, existing studies tend to treat the BNP as an isolated phenomenon, offering little comparison between contemporary BNP support and the previous wave of ERP voting in the 1970s. Our study compares BNP supporters with the only representative sample of NF supporters. Thirdly, the combined role of individual and contextual influences on BNP support has not been examined, despite strong evidence that the BNP focuses on particular local contexts. To address these gaps, we employ multi-level analysis of our survey respondents to analyse the combined influence of individual and contextual factors.

Data and Methods

We examine ERP support at the individual level using survey data gathered and aggregated by Ipsos-MORI from their twice-monthly omnibus survey15 between mid-2002 and mid-2006. We restrict our analysis to England respondents due to the different political environment in Wales and Scotland, where large nationalist parties have operated since the 1970s, and the long-standing focus of the BNP on England. Non-white respondents were also excluded from the analysis given that the BNP has little appeal to non-white voters.16 The resulting sample contains 965 self-identified ERP supporters of the BNP and NF in a combined sample of nearly 150,000 white English respondents aged fifteen or over: the largest sample of extreme right party voters in Britain ever assembled in a survey.17 We use the data to conduct three analyses: first, we examine the social distribution of contemporary extreme right support and compare this with the distribution of NF support revealed in the study by Harrop et al. (1980); second, we examine the attitudes and political priorities of contemporary BNP supporters, and compare these with other party supporters and with those of NF supporters in the 1970s; third, we employ multi-level modelling to examine whether particular constituency contexts are associated with higher levels of BNP support.

Individual-Level Predictors of ERP Support: 1977–8 and 2002–6

In Table 2 we replicate the demographic analysis presented by Harrop et al. using our sample of contemporary ERP supporters. Unfortunately this earlier study did not examine some key variables such as education so these are not included in this initial comparison. Consistent with the wider pan-European literature (e.g. Mudde, 2007, pp. 90–118), the most striking continuity between the two samples is the strong gender bias, with men making up around 70 per cent of the support base of both parties, also suggesting that the BNP's strategy of ‘modernisation’ has not succeeded in increasing the party's appeal among women.

Table 2. Social Distribution of Contemporary BNP and 1970s NF Support (White English Respondents Only)
 % of sample population 2002–6% of BNP support 2002–6% of NF support 1977–8Diff BNP–NFDiff BNP–all 2002 sample
  1. Note:Bold figures indicate groups over-represented among BNP supporters; italics indicate groups under-represented among BNP supporters.

 15–24 years121137−26−2
 25–34 years141316−3−1
 35–54 years343929+10+5
 55 years +403618+18−4
Social class     
 Higher non-manual (AB)22116+5−11
 Lower non-manual (C1)291922−3−10
 Skilled manual (C2)213246−14+11
 Semi-/unskilled manual and residual (DE)273826+12+11
 Greater London11625−19−5
 East Anglia12113+8−1
 East Midlands895+4+1
 West Midlands101423−9+4
Working status     
 Not full-time625532+23−7
 Local authority rented182441−17+6
 Privately rented/other886+30
Social class by age and sex     
 Male, 15–34, ABC16413−9−2
 Male, 15–34, C2DE61725−8+11
 Male, 35 or more, ABC119139+4−6
 Male, 35 or more, C2DE163624+12+20
 Female, 15–34, ABC1723−1−5
 Female, 15–34, C2DE8711−4−1
 Female, 35 or more, ABC12063+3−14
 Female, 35 or more, C2DE181611+5−2

While both samples are male dominated we find significant differences between them on the other major variables. BNP supporters are much older than NF supporters. While 37 per cent of supporters of the earlier NF were under the age of 25, only 11 per cent of contemporary BNP supporters were this young, below their share of the sample population. Overall, 39 per cent of BNP supporters are between the ages of 35 and 54 years old – 10 points higher than in the NF sample – and 36 per cent of BNP supporters are over 55, double the level in the NF sample.18 Several factors may be contributing to this ageing of the extreme right's support base. First, there is evidence of a steep generational decline in racial prejudice in Britain (Ford, 2008) and a similar decline in authoritarian values (Tilley, 2005). This would reduce the potential appeal of the BNP's authoritarian and xenophobic ideology to younger voters. A second possible explanation is a generational effect, namely that these citizens who were socialised during the 1970s – a period characterised by the first wave of ethnic minority migration, salient political conflict over race and immigration, significant support for Powellism and campaigning by the NF – may have retained a greater level of sympathy towards extreme right issues and parties. This cohort would now be in late middle age which is precisely where we see the highest contemporary support for the BNP. A third possibility is that this age gradient is simply one example of the general tendency for older voters to be more politically engaged than the young, a tendency that has become much stronger over the past three elections. Young voters anxious about immigration may be less aware of the BNP, and disinclined to bother voting for it even when they are aware of it.

Turning to social class, as discussed above Harrop et al. and others noted that NF support was concentrated among skilled manual workers, speculating that members of this group are most concerned about immigrant competition. In our contemporary sample, support for the extreme right remains concentrated among the working class, but is now spread relatively evenly across skilled workers, unskilled workers and the residual class of those dependent on state benefits. There is little evidence that those dependent on the state for housing are more likely to turn to the extreme right: BNP supporters in our sample were no more likely to be council tenants than other voters. Aside from age and social class, we also observe a significant shift in the geographical distribution of ERP support. While NF support was concentrated heavily in the West Midlands and Greater London, in contrast the contemporary BNP is far more concentrated in the North of England. While the NF drew 62 per cent of its support from the South of England (i.e. Greater London, South-East, South-West and East Anglia), only 35 per cent of BNP supporters in our sample reside in the southern regions. London, which remains a magnet for immigrants and is Britain's most diverse region, is the BNP's weakest region. The most striking difference in support is found within the North of England region: while the NF was less active in the factory towns of Yorkshire and Lancashire as compared for example to London, these now appear as the strongholds of the contemporary BNP.

Where is the contemporary extreme right drawing its support from in political terms? Unfortunately, we do not have information about the prior voting behaviour of our survey respondents, so we cannot know for sure who is defecting to the BNP. We can, however, compare the social breakdown of ERP supporters with those of other major parties and non-voters (see Table 3). The social distribution of BNP supporters most strongly resembles that of Labour supporters and of non-voters. Labour voters show a similar class, educational and age distribution to BNP supporters, and Labour support is also strongest in the northern regions. Non-voting is concentrated among the least educated and the unskilled manual classes, like BNP support, but it is also far more common among the young than the old, while BNP support shows the opposite tendency. There is very little overlap between the pattern of BNP support and support for the Conservatives and the Liberal Democrats, both of whom tend to draw support from more highly educated, middle-class and southern voters. While we cannot be sure that Labour supporters are indeed defecting to the BNP, both parties are drawing their support from the same social and geographical contexts.

Table 3. Demographic Breakdown of BNP/NF Support and Support for the Other Main Parties
 BNPLabourConsLib DemsUKIPNon-voters
  1. Note:Bold figures indicate over-representation; italics indicate under-representation.

  2. Source:MORI 2002-6; white English respondents aged 18+ only.

 18–34 years252617241440
 35–54 years393730373536
 55 years +363754395124
Social class      
 Professional/managerial (AB)111831311812
 Routine non-manual (C1)192633323123
 Skilled manual (C2)322319182424
 Semi-/unskilled manual and residual (DE)383418192641
 No qualifications343224182632
 GCSE/O level342727223235
 A level111214141211
 South (excl. London)293147475334
 Greater London611911811
 Local authority rented24248121530
 Privately rented/other8869613
N= 149,655      

Attitudes and Priorities of BNP Voters

The surveys that form the basis for this analysis do not ask questions about political attitudes on a regular basis, limiting our ability to probe the attitudinal and policy preferences of BNP voters. However, questions concerning issue priorities, satisfaction with the government and party leaders and opinions about economic prospects are asked regularly enough to compare supporters of the BNP with those of other political parties and non-voters. The issues rated the most important problem facing Britain by different parties' supporters are shown in Table 4. BNP voters are overwhelmingly concerned with immigration: almost 60 per cent of them rate it as Britain's most important problem. The concern of BNP voters with this issue is not unique: immigration is rated as one of the top three problems by supporters of all parties. Yet while other voters divide their attention between about half a dozen issues, BNP supporters focus on immigration almost to the exclusion of all else.

Table 4. Most Important Problem Facing Britain, by Voting Intention (%)
 BNPLabourConsLib DemsUKIPNon-votersFull sample
  1. Notes:White English respondents only. Figures in bold indicate higher than average concern with an issue.

  2. Source:MORI 2002–6.

Defence/foreign affairs/terrorism10241724151821
Crime/law & order51012991111
Others/don't know17282328203427

Table 5 details levels of dissatisfaction with the performance of the government and the leaders of the three major parties, and of pessimism about Britain's economic prospects. BNP voters emerge as a profoundly dissatisfied group of citizens; they are the most negative about the government's performance and the most dissatisfied with the performance of all three major party leaders. This dissatisfaction stretches beyond political performance as BNP supporters are also the most negative about British economic prospects. Nor does it seem to be simply the result of discontent over immigration: BNP supporters who do not rate immigration as the most important problem are no less negative than those who do.

Table 5. Dissatisfaction and Pessimism among Party Supporters (%)
 BNPLabourConsLib DemsUKIPNon-votersFull sample
  1. Notes:White English respondents over 18 only. Figures in bold indicate higher than average dissatisfaction.

  2. Source:IPSOS-MORI political polls, 2002–6.

Dissatisfied with ...
Lab. leader performance85278371865959
Con. leader performance64423151554040
Lib. Dem. leader performance52223114482924
Economy will get worse69275242564841

The intense discontent of BNP voters also distinguishes them from the supporters of their predecessor, the NF. While Harrop et al. (1980, p. 277) likewise found that NF supporters exhibited elevated levels of dissatisfaction with the government (and with the Labour party leadership of James Callaghan), these levels of discontent were lower than those expressed by supporters of the BNP, and were also comparable to those expressed by supporters of the Conservative party. Supporters of the NF were also much less hostile to the Conservative party leadership than contemporary supporters of the BNP. In Harrop et al. (1980), only 50 per cent of NF supporters disapproved of Margaret Thatcher's performance while more than one-third held positive views towards her, suggesting that her efforts to appeal to NF supporters by promising a tougher line on immigration were at least partially successful. The 2005 efforts by Michael Howard to push a tough line on immigration have not had a similar impact on BNP supporters, who are as just as hostile to Howard as to the previous Conservative leaders.19 Howard's failure suggests that the extremely high levels of dissatisfaction among BNP supporters may make them a harder group for the main parties to attract back with ideological or valence appeals.

Demographic and Attitudinal Predictors of BNP Support: Logistic Regression Analysis

Table 6 presents results from a logistic regression analysis modelling ERP support in the 2002–6 MORI data set.20 The multivariate model includes a period trend term to control for the significant rising trend in BNP support since 2002, and a dummy for 2005, the only election year in the sample, when BNP support levels were markedly lower, probably a result of the general tendency for minor party support to fall when a general election is salient.21

Table 6. Demographic Predictors of BNP Support: Logistic Regression Model
 Model 1: Demographics
  • Notes:White English respondents over 18 only.

  • This variable is coded using deviance coding, so the coefficients show deviations from the overall mean.

  • Bold figures denote significant effects:*p < 0.05;**p < 0.01;***p < 0.001.

Survey year (ref: 2002) 
 Linear trend (2002 = 0)0.24***
 2005 (Election year)−0.75***
Age (ref: 18–24) 
 25–34 years0.17
 35–54 years0.32**
 55 years +0.03
Class (ref: AB) 
 Lower non-manual (C1)0.11
 Working class (C2DE)0.76***
Housing tenure (ref: owner-occupier) 
 Rent privately0.11
 Rent from council0.08
 No car in HH−0.33***
 Reads anti-immigrant tabloid0.30***
Education (ref: no qualifications) 
 A level−0.21
Region (ref: London/S.East/S.West) 
 East Anglia0.41***
 East Midlands0.42**
 West Midlands0.69***
 Model fit (pseudo-R square)0.065

The regression model confirms that ERP support in twenty-first century Britain is a male, working-class phenomenon, though once again no difference is found between the three sections of the working class. We therefore include a single dummy for all working-class respondents.22 Individuals from middle age groups are over-represented among BNP supporters in the model, but those aged over 55 are not.23 This provides some tentative evidence that those aged 35–55 years old – who were socialised during the 1960s and 1970s – may have retained a greater propensity to support the extremist fringe. ERP support is also significantly higher in all the Northern and Midlands regions than in London and the South-East. Extreme right support is very rare among those with university education, though lower levels of education have no significant effect. Our model also tests the impact of deprivation. The effects are mixed: unemployed respondents are somewhat more likely to support the BNP, suggesting a greater concern with competition from migrants or minorities. However, those who rent housing from the council are no more likely to support the BNP, while those who do not possess a car are less likely to do so. The latter group, however, may include many such as students or young professionals living in city centres, who are not really ‘deprived’. Unfortunately, more nuanced measures of deprivation were not available in the surveys.

Respondents were also regularly asked which newspapers they read, which allows us to test whether the priming effect of negative media coverage has an impact on British voters. We created a dummy variable for readers of the three main British right-wing tabloid papers noted for hostile coverage of Muslims, immigrants and asylum seekers, namely the Daily Mail, the Daily Express and The Sun. This dummy variable is significant, suggesting that readers of these papers are more likely to support the BNP even after controlling for their social characteristics. While it is possible that this is a selection effect, with xenophobic voters attracted to papers that reflect their views, this finding suggests that high-profile anti-immigrant campaigns such as those appearing in these newspapers could promote support for the extreme right, a finding mirrored in studies of ERP support elsewhere in Europe (Arzheimer, 2009; Ivarsflaten, 2005).

Our initial analysis suggested that BNP voters are attitudinally as well as socially distinctive, expressing very high levels of economic pessimism and political disaffection. Table 7 presents results from logistic models that test the impact of these attitudes. Model 2 includes the political attitudes alone: from this model it is clear that political disaffection, economic pessimism and anxiety about immigration are all strongly correlated with BNP support, with each having a highly significant independent impact. Model 3 tests whether these effects are robust after controlling for the demographic variables from Model 1.24 The combined model confirms that political and social attitudes have a strong impact on BNP support even after controlling for demographic background. The BNP draws support from distinctive social contexts, but its supporters also hold a distinctive pattern of attitudes – highly negative views of the government and main opposition parties, very pessimistic views about economic performance and strong concerns about immigration. While it is not clear from this survey whether these voters agree with the BNP's ideology it is likely that their profound dissatisfaction with the status quo is an important factor in their decision to support the extreme right.

Table 7. Political Attitudes and BNP Support: Logistic Regression Model
 Model 2: Political attitudesModel 3: Political attitudes with demographic controls
  1. Notes:White English respondents only. Bold figures denote significant effects:* p < 0.05; ** p < 0.01; *** p < 0.001.

Political attitudes  
Disapprove of the government1.54***1.70***
Disapprove of Conservative leadership0.49***0.58***
Disapprove of Lib. Dem. leadership1.04***0.90***
Don't know on Lib. Dem. leadership0.68***0.49*
Negative opinion of economic prospects0.57***0.53***
Immigration rated most important problem1.60***1.52***
Model fit (pseudo-R square)0.1250.182

The Local Context of BNP Support: Deprivation and Ethnic Competition

At the individual level our analysis reveals that the BNP is drawing support primarily from politically disaffected middle-aged working-class men in the North of England. Yet the BNP's embrace of intensive grass-roots activism also suggests that local context may be an important factor driving support for the party. Studies by John et al. (2006) and Bowyer (2008) have shown that the BNP has performed most strongly in deprived areas with concentrations of less educated white working-class voters, while Bowyer (2008) has also shown how the presence of a large Muslim minority in the local authority is associated with greater BNP support levels in white working-class wards. These studies, however, suffer from two shortcomings: firstly, they focus on local elections, and it is not clear whether the contextual influences they find will also operate in national elections, where local issues are likely to be less salient; secondly, as these studies focus only on aggregate correlations we do not know for certain whether the aggregate relationships they find actually hold at the individual level. To address these problems we extend the previous logistical regression analysis to incorporate measures of constituency social context.

A two-level logistic regression model is defined, nesting individual respondents within their constituency contexts.25 We apply a simple two-level random-intercept model, which introduces a normally distributed random intercept for each constituency (see Browne, 2005; Rashbash et al., 2008, ch. 9). Maximum likelihood methods can be employed to partition variance in a logistic model, but this involves linearisation techniques which can introduce bias. For this reason, we rejected likelihood methods, and instead estimated the model using Markov Chain Monte Carlo estimation,26 recommended by Harvey Goldstein et al. (2002) as an efficient and unbiased method for partitioning variance in a multi-level logistic model. The estimations were run in the MLWin software programme (see also Browne, 2005). The results of this model are presented in Table 8.

Table 8. Individual and Constituency Predictors of BNP Support: Multi-Level Logistic Regression model
 Model 4: Demographics and constituency context
  1. Note:Bold figures denote significant effects:* p < 0.05; ** p < 0.01; *** p < 0.001.

Survey year (ref: 2002) 
 Linear trend (2002 = 0)0.26***
 2005 (Election year)−0.71***
Age (ref: 18–24) 
 25–34 years0.16
 35–54 years0.33**
 55 years +0.05
CLASS (ref: AB) 
 Lower non-manual (C1)0.04
 Working class (C2DE)0.64***
Housing tenure (ref: owner-occupier) 
 Rent privately0.13
 Rent from council0.09
 No car in HH−0.39***
 Reads anti-immigrant tabloid0.28***
 A level−0.21
Region (ref: London/S.East/S.West) 
 East Anglia0.11
 East Midlands−0.01
 West Midlands0.13
Constituency context 
 % Muslim0.05***
 % non-Muslim Asian−0.03
 % Black−0.06**
 % unqualified0.07***
 % full-time employed0.05***
 Constituency-level intercept variance0.43***

We find strong evidence that constituency social conditions are an important factor in driving BNP support. We find higher BNP support when voters live in constituencies with lower education levels and higher employment rates, even after controlling for individual education and employment status. Consistent with findings in local elections (Bowyer, 2008), we find that the presence of a large Muslim community at constituency level is associated with higher levels of BNP support. In contrast, the presence of non-Muslim Asians has no significant effect while BNP support is actually lower in areas with larger black populations.27 This suggests that the appeal of the modern BNP is more subtle than the crude racism and xenophobia of the old NF, whose support was correlated with the presence of any non-white ethnic group. Unlike earlier years, Indian and black Caribbean minorities are a more established and accepted part of society (Ford, 2008) and do not excite as much white hostility. Meanwhile, and like European ERPs more generally, the BNP has abandoned biological racism, shifting its programmatic emphasis instead towards white anxieties over the more socially and geographically segregated Muslim minorities who can more easily be stigmatised as hostile to British values and supportive of terrorism.28

The substantive impact of these variables is illustrated in Table 9. Here we show the predicted levels of support for the BNP at low and high values for each of the predictor variables, as well as the combined impact of each set of predictor variables on BNP support levels.29 Many of the predictors of BNP support are associated with a doubling or more of BNP support, albeit from a very low baseline of 0.5 per cent or less. Gender and education have the strongest impact out of the demographic variables, each more than doubling BNP support. Combining variables, we find that our model predicts nearly 3 per cent of BNP support among the middle-aged, less well-educated working-class tabloid-reading men who form the demographic core of the party's support in our model. Attitudinal variables have effects of broadly similar magnitude, with the biggest effects associated with prioritising immigration and being dissatisfied with the government. The combined impact of the attitudinal predictors is dramatic: one in eight of voters who are disaffected with all the main parties, negative about the economy and prioritising immigration are predicted to support the BNP. The constituency context variables also have effects of broadly similar magnitude, with most being associated with approximately a doubling in BNP support. The impact of low education levels in the local area, however, is considerably stronger. Moving from a constituency with very low numbers of unqualified voters to one with very high levels is associated with a quadrupling in BNP support from 0.4 per cent to 1.7 per cent. The cumulative impact of these variables is again dramatic: 6.5 per cent of voters living in the strongest BNP contexts, with large Muslim and African populations, low education levels and high employment levels, are predicted to support the party, compared with 0.1 per cent in the weakest contexts. This simulated effect accords closely with actual levels of BNP support in the 2005 general election. In the five constituencies30 in our sample with the most ‘BNP favourable’ social contexts – large Muslim populations, low education and employment levels and small black populations – the BNP won an average of 7.1 per cent of the vote.

Table 9. Simulated Impact of Demographic, Attitudinal and Contextual Factors on Predicted BNP Support
Predicted BNP support (%)Support lowSupport highChange
  1. Note:5th percentile compared with 95th percentile.

 Age (55 plus/35–54)0.690.900.21
 Working class (no/yes)0.561.040.48
 Education (degree/no quals)0.320.900.58
 Sex (female vs. male)0.491.250.76
 Read anti-imm. tabloid (no/yes)0.690.910.22
 Combined demographics0.132.892.76
 Govt satisfaction (sat./dis.)0.180.930.75
 Con. leader satisfaction (sat./dis.)0.410.730.32
 Lib. Dem. leader satisfaction (sat./dis.)0.400.990.59
 Economic prospects (pos./neg.)0.400.680.28
 Imm. most imp. prob. (no/yes)0.391.771.38
 Combined attitudes0.0812.3012.22
Constituency context   
 % Muslim0.641.060.42
 % Black0.800.48−0.32
 % No qualifications0.381.671.29
 % in FT employment0.531.090.56
 Combined context0.106.576.47


In his seminal study of the 1970s NF, Christopher Husbands (1983, p. 147) warned against an interpretation of extreme right support in Britain that subscribed to Hofstadter's epigram: ‘Third parties are like bees: once they have stung, they die’. Rather, Husbands drew attention to the way in which – in some local districts – the NF successfully mobilised support by drawing on sensitivities and susceptibilities deeply entrenched in working-class culture. Observing trends in NF support, Husbands found it difficult to escape the judgement that a successor organisation to the NF, one with a similar appeal but which eschewed the Front's political ineptitude, could achieve a greater and enduring level of electoral success.

Consistent with this prediction the early years of the twenty-first century have seen unprecedented levels of support for the latest representative of the extreme right, the BNP. Our analysis of survey evidence suggests that the BNP's efforts to ‘modernise’ and offer a more moderate and democratic ideology are winning over a small but fast-growing band of new recruits. The new extreme right support base has a clear demographic and attitudinal profile – white, middle-aged, working-class men who are extremely negative about mainstream politics and very worried about immigration. BNP supporters cluster where education levels are low and where significant Muslim communities live.31 As others observe, the rapid growth of a poorly resourced new party with a very negative public image in a predominantly first-past-the post system with high barriers to entry is an extraordinary phenomenon (John and Margetts, 2009, p. 505). At the outset, we considered three ways in which such a phenomenon could be explained: failures of political management, the emergence of a new ideological division and sociological change. Here we speculate on the possible impact of these three factors in light of the evidence we have presented, and consider their implications for the future development of the BNP.

With regard to political management our evidence strongly suggests that the persistent inability of the mainstream parties to provide a convincing response to voter anxiety about immigration is an important factor contributing to the emergence of the BNP. BNP voters in our sample overwhelmingly cite immigration as the most important problem facing the country, and they are far from alone in their concern: 16 per cent of all the voters in our sample rated immigration as the most important national issue. While the experience of the 1970s suggests it is possible for a mainstream party to win credibility on the issue of immigration – Thatcher's promise of an ‘end to immigration’ certainly contributed to the subsequent collapse in NF support – winning such credibility now will be much more difficult than it was 30 years ago. Given that the bulk of current migrants to Britain arrive through family reunion, asylum or settlement from the EU, contemporary British governments have only limited discretionary control over migration. All these forms of migration are protected by law or international treaty, and no British government can easily prevent them. Even when governments do succeed in passing restrictive reforms – as Labour has with the asylum system – such changes may not alter public perceptions. The anti-immigration media has helped drive widespread public distrust of the official migration figures, and many voters may conclude that rising local diversity is the result of ‘uncontrolled’ migration when in fact it is the product of natural increase in British ethnic minority populations. Thirdly, pushing ‘tough’ policies on immigration and asylum in a country where one voter in ten will soon be non-white comes with substantial political risks. Policies that are perceived as discriminatory or pandering to xenophobia may alienate ethnic minority voters and young well-educated white voters. It could also be a bad electoral bargain: the Conservatives would not improve their prospects of a Westminster majority by appealing to working-class voters in Labour-dominated urban industrial constituencies they have no realistic chance of winning. All of these factors suggest that the perceived level of immigration is likely to remain unacceptably high in the eyes of many voters, and the main political parties are unlikely to run serious and sustained campaigns to halt it. Thus the opportunity for the BNP to win over valence voters anxious about immigration is likely to persist.

The contextual analysis presented here also suggests that the BNP – like other European ERPs – is successfully exploiting perceptions of conflict between the white majority and Muslim minorities (e.g. Cofféet al., 2007). The party's ability to mobilise white hostility to Muslims raises the possibility that identity conflicts could displace traditional left–right politics in diversifying local communities. Muslim Britons are the most economically deprived ethnic group and therefore have similar economic interests to the white working class. However, negative stereotypes, religious and cultural differences and social and geographical segregation may easily displace perceptions of common interest and result in an ‘us vs. them’ mentality. White working-class voters may desert Labour if they perceive it to be ‘the Muslims' party’32 just as in many parts of America white working-class voters deserted the Democrats because they were perceived as representing black interests (Huckfeld and Kohfeld, 1989). The BNP would not be the only possible destination for such voters, but it is an obvious one and unlike its predecessors is actively experimenting with ways to appeal to them. As levels of diversity increase, these cross-cutting identity conflicts may become more common, complicating efforts to win support for redistributionist and welfare policies as white voters become unwilling to support state assistance for those ethnically and religiously different to themselves (see Banting and Kymlicka, 2006 for a review).

The BNP's emergence poses a challenge chiefly to the Labour party. While our survey evidence cannot demonstrate that the BNP is directly recruiting Labour voters, the social profiles of BNP and Labour supporters are strikingly similar, and BNP support is also highly concentrated in areas dominated by Labour. For instance, 33 of the 34 constituencies where the BNP polled above 5 per cent in 2005 have Labour incumbents. These constituencies have been electing Labour MPs continuously for an average of nearly 40 years. At the local level, the BNP is most powerful in authorities such as Stoke and Barking which were dominated by Labour for decades after 1945. In these authorities, the rise of the BNP has gone hand in hand with the decline of the ruling Labour party (Wilks-Heeg, 2009). Several factors are at work here. There has been a steady erosion in the cultural and institutional links that once bound white working-class voters to Labour, such as large industrial employers, trade unions, working men's clubs and local Labour branches, thereby opening these areas to political competition (indeed the BNP recently launched its own trade union). The Labour-dominated councils in many of these areas also often provided poor governance and became associated with the economic decline and rising ethnic diversity that occurred ‘on their watch’. Finally, the other Westminster parties have virtually no organised presence in many of these areas, so the BNP faces very little competition for the votes of those unhappy with the local Labour party.

All three of these factors run deeper than the moral panic over migration to which the BNP's success is commonly attributed. Our findings suggest that the BNP has succeeded in establishing itself as a credible alternative for working-class voters angry about immigration, threatened by Muslims and hostile to the political establishment. The underlying trends that have driven these voters to the BNP are set to continue, and as such the BNP looks likely to remain a significant force in British politics. The party that should be most concerned about this development is Labour. We do not know the former loyalties of BNP voters: some, as qualitative research suggests, may be working-class ‘Old Labour’ voters who feel abandoned by a party that no longer reflects their values (Goodwin, forthcoming [b]); others may be working-class Tories who voted for Thatcher but are unimpressed by Cameron; still others may previously have been too hostile to mainstream politics to vote for any party. However, it is clear from our data and from the electoral returns of the past five years that the BNP has prospered in areas Labour once dominated. For the first time in decades, Labour now faces a serious and organised threat to its control of these council seats and constituencies. Developing an effective response to this threat at the local level will be a crucial challenge for the party in the coming years.

About the Authors

Robert Ford is Hallsworth Research Fellow in the Institute of Social Change, University of Manchester. His research focuses on racial attitudes, immigration and political choice.

Robert Ford, Institute of Social Change, University of Manchester, Humanities Building, Bridgeford Street, Oxford Road, Manchester M13 9PL, UK; email:

Matthew J. Goodwin is ESRC Postdoctoral Research Fellow at the University of Manchester where his research focuses on political behaviour and extremist political organisations. He is co-editor of The New Extremism in Twenty-First Century Britain (Routledge, 2010, forthcoming) and author of a forthcoming book on the British National party.

Matthew J. Goodwin, Institute of Political and Economic Governance, University of Manchester, Humanities Building, Bridgeford Street, Oxford Road, Manchester M13 9PL, UK; email:


  • 1

    Prior to the BNP a short-lived splinter group from the NF, the National party (NP), elected two councillors in 1976 in Blackburn; the British Fascist (BF) movement elected two candidates in 1926 in Lincolnshire and Oswald Mosley's British Union of Fascists (BUF) elected one candidate in Suffolk in 1938. In 1993, the BNP elected one local councillor in Millwall ward, Tower Hamlets.

  • 2

    On ‘no-to-elections’ see Wells (1991).

  • 3

    British Nationalist, July 1983, p. 4. In 1987, two unofficial candidates represented the BNP in two constituencies in south London.

  • 4

    Spearhead, 1992, no. 281, p. 9.

  • 5

    BNP, Language and Concepts Discipline Manual. Available from:[accessed 1 February 2009].

  • 6

    British Nationalist, July 1998, p. 4.

  • 7

    British Nationalist, January 1998, p. 1. The party claimed that ‘teams’ of activists were distributing these leaflets though it is difficult to know how many were distributed. See British Nationalist, February 1998, p. 6.

  • 8

    BNP, Statement of Accounts – Regional Accounting Unit, 2005, p. 1. Available from:[accessed 27 July 2009].

  • 9

    In adopting a ‘working-class authoritarianism’ approach Whiteley drew on work by Lipset (1969) and Lipset and Raab (1971), who argued that working-class radical right support was the result of low education, economic insecurity, authoritarian family patterns and the status anxiety of whites faced with the challenge of aspirant blacks.

  • 10

    As Husbands (1983, p. 34) notes, ‘The concept of “political tradition” is evoked to accommodate the fact that NF voting often either contained an episodic history of race-related mobilisation (one that might go back to before the turn of the century) or else took advantage of a specific and distinctive local political culture’.

  • 11

    The authors asked NOP Market Research to cumulate 22 studies of voting intentions conducted between 6 October 1977 and 5 April 1978 as part of its regular random (weekly) omnibus surveys. Details of the survey methodology are provided in Harrop et al. (1980, p. 273).

  • 12

    Research by Husbands (1983), however, found a different pattern. In his study elevated NF support was found among both skilled and unskilled working-class men. Husbands found the strongest NF support among both the youngest (25 and under) and the middle-aged (46–55).

  • 13

    John et al. (2006, p. 16) also observe a tipping point in that when Asian groups constitute up to 7 per cent of the population in a council area electoral support for the BNP increases with the percentage of Asians in the local population while above this level support for the extreme right actually decreases.

  • 14

    Perceptions of ethnic competition and threat, in particular linked to Islam and Muslim communities, also figure prominently in the motivational accounts of BNP activists (Goodwin, forthcoming [b]).

  • 15

    The survey uses a nationally representative quota sample. There are 210 sampling points, each consisting of a census ward or ward-sized area of approximately 3,500–8,000 households. Each sampling point is carefully selected to ensure that the sample universe is representative at regional level on a large number of demographic and other criteria. Within each sampling point, an interviewer is set a quota of ten interviews, with a different quota for each sampling point based on its census profile – quotas are set on gender, age, housing tenure and work status (full-time workers vs. others). Respondents were asked, ‘How would you vote if there were a general election tomorrow?’ and then shown a card listing ‘Conservative’, ‘Labour’, ‘Liberal Democrat’ and ‘other’. All those who replied that they were undecided or refused to answer were then asked: ‘Which party are you most inclined to support?’ Voting intention for each party combines those who expressed support at either the first or second question. Self-identified supporters of the extreme right were those respondents whose verbatim answer recorded support for ‘British National party’, ‘BNP’, ‘National Front’, ‘NF’ or any obvious variant of those. The vast majority of included respondents (around 90 per cent) volunteered a BNP affiliation. MORI have suggested that much of the remaining minority may be accounted for by respondents' tendency to misremember the names of political parties (MORI, personal communication to authors, April 2009).

  • 16

    The lack of appeal the BNP holds for ethnic minority Britons is clear in our sample: not one of the over 12,000 ethnic minority respondents volunteered support for the party.

  • 17

    Respondents aged 15–17 were retained in our analysis as they would be old enough to vote at the next general election. Replications of our analyses with these individuals excluded had no impact on the results. The sample is likely, however, to suffer from some social desirability bias. Citizens are often less willing to admit to supporting extremist parties such as the BNP ‘because they are aware that mainstream society views such support as a violation of the democratic norm’ (Knigge, 1998, p. 262; see also Eatwell, 2000, pp. 175–6). Our sample is therefore likely to be an underestimate of true BNP support, and to over-represent the most committed BNP partisans. It is impossible for us to estimate the extent of the bias this introduces and we can only suggest that the robustness of our findings be tested in future when surveys of BNP support employing less obtrusive methods (for example, internet-based surveys) become available.

  • 18

    While the white English population has aged somewhat between the late 1970s and the early 2000s, this demographic change is much smaller than the change in age distribution seen here: the proportion of Britons over 50 increased by about 3 percentage points between 1971 and 2001 (Smith et al., 2005).

  • 19

    Although one factor that may be relevant though not researched is Howard's Jewish heritage which is frequently referenced in BNP literature.

  • 20

    Unfortunately, we have been unable to obtain the Harrop et al. data in order to conduct a comparison regression analysis on previous NF support.

  • 21

    Campaigns have commonly been found to reactivate voters' latent partisan preferences in this fashion. Publicly expressing support for the BNP may also attract more social sanction in the run-up to an election, when such expressions of support will be taken more seriously by others.

  • 22

    Likelihood ratio tests confirm that a model with a single dummy for working class fits as well as one with three coefficients for each working-class group (C2, D and E), so we opt for a single coefficient for parsimony.

  • 23

    Age effects were also examined using a continuous measure of age which was unfortunately only available for half of the survey sample. The inclusion of a squared term confirmed non-linearity in the age effects, with peak support found at around age 40.

  • 24

    The effects of these variables are generally unchanged in the combined model, though significance levels are in many cases reduced by the much smaller sample size. They are not shown for reasons of space. Full coefficients are available from the authors on request.

  • 25

    This model is defined only for the demographic model, as the number of BNP supporters in each constituency is too low to support multi-level analysis of the sub-set of respondents who answered the political attitudes questions.

  • 26

    Iterations were run after a ‘burn-in’ period of 500 iterations to allow model convergence. These are omitted from the sample used to calculate model statistics.

  • 27

    We also tested more disaggregated models of ethnic context, which suggested that the positive effect of Muslim presence is particularly strong for Pakistani Muslims, concentrated in the industrial regions of the North of England where they are a highly visible and segregated minority. Bangladeshi Muslims, who concentrate instead in very diverse areas of inner London where they are a less visible source of threat, are less associated with BNP support.

  • 28

    We tested the robustness of these findings by also conducting analyses of constituency predictors of BNP support in the 2001 and 2005 general elections. We conducted three analyses – a logistic regression analysis of whether the BNP stood a candidate, an OLS analysis of support levels where they did and a Tobit regression which attempts to model both processes together. In all three cases, we found that the pattern of contextual effects was very similar to those in the MORI sample.

  • 29

    All effects are calculated in the full model with the other predictors set at their mean values.

  • 30

    These five constituencies are: Blackburn (BNP 5.4 per cent), Bradford North (6.0 per cent), Rochdale (4.3 per cent), Dewsbury (13.1 per cent) and Batley and Spen (6.8 per cent). All five have Labour incumbent MPs.

  • 31

    This profile could of course change as the BNP's support rises. However, at present there are no individual-level data to test this possibility.

  • 32

    Rhodes (2006) shows how precisely such identity conflicts fuelled the BNP's rise in Burnley.