By Michael Biggs and Steven Knauss
The BNP’s membership list was leaked in November 2008 by a disgruntled activist who had been expelled late in 2007; he has since admitted responsibility and been convicted. The BNP never challenged the list’s authenticity, merely stating that it was out of date. The list is apparently a complete record of membership at November–December 2007. Of the 13,009 individuals listed, 30 were missing a current address, 138 had a foreign address, and 41 lived in Northern Ireland. Of the remaining members, 12,536 (97.9 per cent) can be precisely located in Britain using the postcode field of their address (Office of National Statistics, 2004, 2008). Postcodes provide exceptionally fine resolution, down to the street level.
The distribution of members diverges significantly from the distribution of voters. The correlation of votes with membership, across the 628 constituencies in Britain, is surprisingly modest (r ¼ 0.46). The party contested only one in five seats, but the correlation is scarcely higher in those alone. Voting also gives a misleading impression of the national distribution of the party’s support. Wales and Scotland provide over three times the proportion of members compared with voters.
Members must be matched with a population denominator. Data come from the 2001 Census, conducted in April. The great majority of members on the leaked list had joined since this date, as the BNP had 2,173 members in November 2001 (Copsey, 2008: 137). The BNP recruited only ‘indigenous Caucasian’ people (Copsey, 2008: 238). We count adults who defined their ethnicity as ‘White British’, including ‘White Scottish’. The proportion of white British adults belonging to the BNP was 0.032 per cent across Britain.
For statistical analysis, we use the finest geographical unit defined by the Census, the ‘output area’. This is a very small neighbourhood; the median covers an area of 6 hectares and contains 280 people. There are 218,038 neighbourhoods (as they will be termed) in Britain: the BNP was present in 10,165 (4.7 per cent) of them. Most of those had a single member; 11 was the maximum. The highest proportion was 5.7 per cent.
We begin with independent variables capturing economic insecurity. These are measured ecologically, as the fraction of people in the neighbourhood with a particular characteristic, though they are proxies for individual characteristics predicting support for the BNP. Education is divided into three categories: no qualifications, qualifications below university degree, and degree (denominated by people aged 16–74 years). Class is divided into five categories, from routine and semi-routine to managerial and professional (denominated by occupied population). The unemployment rate is also measured (denominated by the economically active). Alongside these sociological staples, housing is included because the BNP promotes the myth that foreigners are given privileged access to public housing. Housing tenure is divided into three categories: owned or mortgaged, rented from the local authority, and private rental (including other arrangements). Overcrowding, as defined by the Census, is also measured. (In both cases the denominator is households.) We expect, then, that white British adults are more likely to belong to the BNP in neighbourhoods with lower education, lower social class, higher unemployment, more private renting, and greater overcrowding. Control variables are entered to reflect findings that BNP voters are disproportionately male and middle aged (Ford and Goodwin, 2010; Cutts et al., 2011). Additional controls are population density and the proportion of people living in communal establishments like prisons.
For Hypotheses 1–3, we defin