Localised variations in South Asian turnout: a study using marked electoral registers Ed Fieldhouse With David Cutts, Paul Norman, Mark Tranmer and Kingsley Purdam
Turnout: Secular Decline vs Trendless Fluctuation Percentage Turnout at General Elections as proportion of registered electorate 1929-2001 90 80 70 60 50 40 30 20 10 0 1929 1931 1935 1945 1950 1951 1955 1959 1964 1966 1970 1974f 1974o 1979 1983 1987 1992 1997 2001
Reported Turnout (MORI 2001) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Q.1. Some people did not get along to vote at the General Election on the 7th June. Did you vote at the General Election on the 7th June? Yes No Non-ethnic minority Ethnic Minority Black Asian
Reported turnout by ethnic minority status 100.0 90.0 80.0 70.0 60.0 50.0 40.0 30.0 20.0 10.0 0.0 Austria Belgium Switzerland Czech Republic Germany Denmark Spain Finland Great Britain Greece Hungary Ireland Israel Italy Non-minority Luxembourg Holland Norway Poland Portugal Sweden Ethnic minority Source: European Social Survey
Reported Registration BES 86% of ethnic minorities registered to vote, 95% of whites 1997 BES allowed more detailed breakdowns (highest amongst Indian and white population) ONS 1991 White 93%, Asian 85%, Black 76% MORI survey 2001: Asian, 92%, Black, 90%, White, 97% Over 15% of MORI 2001 black respondents quoted this as reason for non-voting (all e.m. 10%, white, 6%)
Explaining Ethnic Differences Generic factors Younger age profile of ethnic minority communities Social class Higher levels of social and economic deprivation Systemic effects: e.g. geography and marginality Community Specific Factors Community engagement Representation in political institutions and parties Relevance of mainstream politics Exclusion and racism Neighbourhood effects?
Neighbourhood effects: possible explanations (according to Johnston, earlier today) Social Interaction Environmental Selection (or residential segregation?) Emulation Environmental observation Local pressure And Unobserved individual effects.
Distinguishing contextual and compositional effects Macro analysis estimates for population characteristics will contain contextual as well as compositional effects, but. Micro analysis area residuals will include unobserved compositional as well as contextual effects Evidence for neighbourhood effects is circumstantial Imperfectly specified models Unobserved heterogeneity
Why should neighbourhoods effect turnout? Effects on voting Applicability Social Interaction - social capital and embededness in local community Environmental Selection - resources Emulation - system benefits or civic duty and social norms Environmental observation - collective benefits Local pressure - mobilisation Unobserved effects - neighbourhood clustering is an artefact
Why should neighbourhood effects be important for South Asian electors? Polarisation in turnout (Indian and Pakistani/Bangladeshi) Spatial segregation/parallel lives But. Cohesive communities: talk together and vote together? Community leadership and mobilisation More powerful household effects in large households (and household clustering)
New study for Joseph Rowntree Foundation Registration and turnout Mis-reporting Survey non response Marked electoral registers Sample of 100 wards and two case study metropolitan boroughs 2001 Census 2001 (April 29 th 2001) General Election (June 7 th 2001) Nam Pehchan
Sample design % South Asian Pop. No of Wards Sample Total Pop. South Asian Pop. Excluded 0% 2057 0 N\A 0 Group A >0% - <0.5% 5134 20 140,030 314 Group B 0.5% - <2% 1972 20 125,955 1,495 Group C 2% - <10% 1025 20 158,849 6,721 Group D 10% - <20% 201 20 187,869 27,669 Group E 20% + 163 20 225,984 85,372 Total 10552 100 838,687 121,571 Note. All registered electors in the sampled wards are included, though this will be somewhat less than the total pop (column 3), which includes all persons of all ages.
Burngreave location
Burngreave context Name of constituency Major Census Region (Britain) Sheffield, Central Yorkshire and Humberside County South Yorkshire Winner of seat 2001 Lab win Party in 2nd place 2001 LibDem 2nd % Turnout 2001 48 % Majority 2001 42 % Vote Conservative 2001 11 % Vote Labour 2001 61 % Vote LibDem 2001 20 Rank index of multiple deprivation 2000: 60 Constituency Sheffield Central MP - Richard Caborn Turnout 2001 48%
Burngreave: Ethnic composition, 2001 Resident Burngreave Sheffield England White population 58.5 91.2 90.9 of which White Irish (percentage) 0.9 0.6 1.3 Mixed 4.8 1.6 1.3 Asian or Asian British 22.9 4.6 4.6 Indian 0.8 0.6 2.1 Pakistani 18.8 3.1 1.4 Bangladeshi 0.6 0.4 0.6 Other Asian 2.6 0.5 0.5 Black or Black British 12.2 1.8 2.1 Caribbean 6.3 1.0 1.1 African 5.2 0.6 1.0 Other Black 0.8 0.1 0.2 Chinese or Other Ethnic Group 1.6 0.8 0.9 Burngreave Sheffield England and Wales Christian 48.0 68.6 71.8 Buddhist 0.2 0.2 0.3 Hindu 0.4 0.3 1.1 Jewish 0.1 0.2 0.5 Muslim 27.6 4.6 3.0 Sikh 0.2 0.2 0.6 Other religions 0.4 0.2 0.3 No religion 13.8 17.9 14.8 Religion not stated 9.4 7.8 7.7
NAM Pechan Results for Sheffield Burngreave Religion Common Hindu Muslim Sikh Buddhist Other Clashes Total Language Bengali 0 0 0 0 0 0 0 0 0.00% Gujerati 0 0 0 0 0 0 0 0 0.00% Hindi 0 1 0 0 0 0 0 1 0.10% Punjabi 0 0 3 6 0 0 1 10 0.60% Pushtu 0 0 0 0 0 0 0 0 0.00% Singhalese 0 0 0 0 0 0 0 0 0.00% Sylheti 0 0 0 0 0 0 0 0 0.00% Tamil 0 0 0 0 0 0 0 0 0.00% Urdu 0 0 460 0 0 0 0 460 27.30% Other 0 0 0 0 0 1 0 1 0.10% Muslim Languages 22 0 1031 2 0 0 1 1056 62.70% Hindu Languages 44 9 0 1 0 0 1 55 3.30% Common 39 0 28 0 0 0 0 67 4.00% Clashes 7 1 21 1 0 0 5 35 2.10% Total 112 11 1543 10 0 1 8 1685 100.00% 6.60% 0.70% 91.60% 0.60% 0.00% 0.10% 0.50% 100.00% Note: 20% of register coded as South Asian
Registration and turnout by South Asian and all other electors, Burngreave, 2001 Not South Asian South Asian Total Population Over 18 8296 1852 (18%) 10148 All registered 6891 1685 (20%) 8576 Registered eligible 6481 1649 (20%) 8130 Voted 2962 810 (21%) 3772 % Registration 83 91 85 % Turnout (of eligible 46 49 46 electorate) % Turnout (of pop 18+) 36 44 37
Estimated registration by OA
Variation in turnout by OA (all electors)
South Asian Electors %
Carstairs index of deprivation
South Asian turnout
South Asian turnout v electorate share South Asian turnout 100 90 80 70 60 50 40 30 20 10 0 0 20 40 60 80 % South Asian Electors
Turnout by other ethnic groups
South Asian and other turnout % turnout 120 100 80 60 40 20 0 1 5 9 13 17 21 25 29 33 37 41 45 OA ranked by overall turnout Votesa% othervote%
Two level model (person and OA)
Residuals from model
Two level model with South Asian % and interaction
Two level model with compositional variables
3 level model (person, household, OA)
3 level model with compositional variables
Conclusions Turnout might be expected to influenced by neighbourhood context (but some variation might be due to residential selection coupled with unmeasured effects) Neighbourhood effects might be expected to be stronger for South Asian groups due to community cohesion and household size South Asian turnout is (probably) higher then that of other ethnic groups Impact of ethnic origin varies across areas Preliminary evidence from one area shows size of South Asian population is key indicator in determining whether South Asian turnout is higher Homeownership, % of single pensioners and deprivation also impact on turnout levels. Household variation much more substantial than OA level variation