Population turnover for deprived neighbourhoods: structural or pathological?

Transcription

1 Population turnover for deprived neighbourhoods: structural or pathological? BSPS Conference, Canterbury Sept 2005 Nick Bailey & Mark Livingston Scottish Centre for Research on Social Justice University of Glasgow Contact:

2 Context Spatial segregation is a source of injustice and this is recognised by current Govt My vision is of a nation where no-one is seriously disadvantaged by where they live (PM s Foreword to SEU, 2001) "By 2021 no one should be seriously disadvantaged by where they live. (PMSU, 2005) We are committed to building a better Scotland, where a child s potential, and not their postcode, will decide their future. (Scottish Executive, 2002)

3 Context Migration a dynamic view of area deprivation Net flows (preventing) area change Gross flows and area instability Geography of flows and connection/isolation

4 Gross flows and area deprivation Gross turnover for SOAs in England by deprivation Gross turnover ratio 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% Over 1-to-4 Up to 1-to-4 Up to 1-to-5 Up to 1-to-6 Less than 1-to- 0% Least depvd Most depvd Deprivation decile

5 Gross flows and area deprivation High turnover seen as a characteristic of deprived areas which contributes to their problems Empty properties, higher management costs, opportunity for vandalism Erosion of social networks, loss of support Loss of social control, problems of disorder Associated with range of problems, e.g. mental health problems (Faris and Dunham, 1939; Silver et al, 2002)

6 Gross flows and area deprivation Why do deprived areas have higher levels of turnover than non-deprived? Neighbourhead-level explanations Area pathologies Structural explanations Compositional factors (Rossi, 1980) Wider context city-regional, regional (Dielman et al, 2000) Aim to separate out these various factors to inform policy responses

7 Individual migration Determinants of individual migration Age (life stage) Ethnicity Household type and size Employment status Education Housing tenure Health Caring responsibilities Student SARs data, logistic regression model (Bailey and Livingston, 2005)

8 Family ty pe Nos. caring in h h ld Nos. LLTI Eco. in status hhldtenure of FRP quals. Educ. Age bandssex Ethnicity Constant Student (not) LP w. nondep only LP w. dep/nondep Cple. nondep only Cple w. dep/nondep Couple Single (default) 0 (default) 12 Lev/Grp 4+ Lev/Grp 3 Lev/Grp 2 No quals, Lev/Grp 1 Inactive Unempld Employed (default) Private renting Social housing Ow ner occ. (default) 2+ 0 (default) 1 Other Mixed Chinese Black/Brit Asian/Brit White (default) Female (Male) (default) Determinants of individual migration Regression coefficient (B)

9 Measuring gross turnover

10 Area turnover Determinants of area churn Age composition (life stage) Area ethnicity Household type and size Area employment status Area education (age stdised no quals) Housing tenure Area health (age stdised LLTI) Area caring responsibilities Area students CAS data for SOAs/DZs, linear regression No communal establishment

11 Area Determinants of Churn status (16-74) Care Family Stud Net Age Ethnicity Tenure Negative net turnover Positive net turnover Proportion of students squared Percentage students % Households w ith non dependent children % Households w ith dependent children % Single/Couples w ith out children % 50 or more hours % hours % 1-19 hours No Caring Age standardised no qualifications % Unemployed % Inactive % Employed % Private renting % Social housing % Ow ner occupier Age adjusted limiting long term illness % Other % Mixed % Chinese % Black % Asian % White % 75over % % % % % % % % 5-14 % 0-4 justed R squared = Regression Coefficient (Standardised)

12 Determinants of Area Churn including deprivation Economic status Age Ethnicity Deprivation Tenure (16-74) Care FamilyStudNet Negative net turnover Positive net turnover Proportion of students squared Percentage students % Households w ith non dependent children % Households w ith dependent children % Single/Couples w ith out children % 50 or more hours % hours % 1-19 hours % No Care Age standardised no qualifications % Unemployed % Inactive % Employed % Private renting % Social housing % Ow ner Occupier Age adjusted limiting long term illness 100% (most deprived) 98-99% 96-97% 91-95% 86-90% 71-85% 50-70% 30-49% 15-29% 10-14% 6-9% 4-5% 2-3% 1% (Most Affluent) % Other % Mixed % Chinese % Black % Asian % White % 75over % % % % % % % % 5-14 % 0-4 Adjusted R squared = Regression Coefficient (standardised)

13 Deprivation Section 100% (most deprived) 98-99% 96-97% 91-95% 86-90% 71-85% Deprivation 50-70% 30-49% 15-29% 10-14% 6-9% 4-5% 2-3% 1% (Most Affluent) Regression coeffecient (standardised)

14 Model including house building and density Economic status Age Ethnicity Deprivation Tenure (16-74) Care FamilyStudNet Log of density Locality Housebuilding rates year 5 Negative net turnover Positive net turnover Proportion of students squared Percentage students % Households w ith non dependent children % Households w ith dependent children % Single/Couples w ith out children % 50 or more hours % hours % 1-19 hours % No Care Age standardised no qualifications % Unemployed % Inactive % Employed % Private renting % Social housing % Ow ner Occupier Age adjusted limiting long term illness 100% (most deprived) 98-99% 96-97% 91-95% 86-90% 71-85% 50-70% 30-49% 15-29% 10-14% 6-9% 4-5% 2-3% 1% (Most Affluent) % Other % Mixed % Chinese % Black % Asian % White % 75over % % % % % % % % 5-14 % 0-4 Adjusted R squared = Regression Coefficients (Standardised)

15 Summary Compositional factors most important in relation to churn (86% of variation) High deprivation associated with slightly higher levels of churn but so is low deprivation House building rates in locality associated with higher churn even after removing neighbourhood-level population growth Explore residuals for further clues to drivers of churn

16

17

18

19

20 Conclusions Compositional factors explain most variance in churn Some additional neighbourhood-level effect from deprivation but fairly weak Errors partly systematic suggesting further contextual, compositional and wider-area factors at work Area model adds to analysis of individual data

21 Policy implications Conclusions Challenges views which pathologise deprived areas by highlighting structural basis of high churn in deprived areas Provides basis for distinguishing between churn which is structural and that which does reflect local problems Supports arguments for more mixed communities but mix needs to be seen in terms of demographics (age, household types) not incomes