SLIDE 1
High-unemployment neighbourhoods in weak labour markets
”The socio-political challenges of medium-sized cities” 10 December 2010, Keele University Alex Fenton (aff28@cam.ac.uk) CCHPR, Department of Land Economy
SLIDE 2 Spatial disparities in unemployment - Two modes of analysis “Regional economics”
Primarily economics-based A puzzle for classical economics?
Large spatial scale
Cities, regions
Non-housing explanations
Agglomeration Infrastructure Human capital Global competition
“Neighbourhood geography”
Variety of disciplines
Small spatial scale
Neighbourhoods, estates, districts
Housing explanations
Sorting by allocation & subsidy Sorting by stock & price Domestic and international
migration
Neighbourhood effects?
SLIDE 3 Background questions If there are inequalities, do they matter?
Thresholds / non-linear effects of concentrated disadvantage “Cultures of worklessness”? “Equality”
And if so, what should we do about it?
Individual-level intervention (coercive / supportive) Broad redistribution Neighbourhood-level ABIs Sub-regional economic development
SLIDE 4 Three empirical studies of high-unemployment neighbourhoods 25-year estimates of neighbourhood unemployment rates in England & Wales
Joseph Rowntree Foundation, “Communities in recession”, 2009
Modelling and cluster analysis of employment-deprived neighbourhoods in England
Communities and Local Government, “Typologies of Place”, 2009-10
“Why do neighbourhoods stay poor?” - mixed methods study of Birmingham poverty neighbourhoods
Supported by Barrow Cadbury Trust, 2008-10
SLIDE 5 Project 1: Neighbourhood unemployment rates 1985 - 2010 JSA claims data + GIS-derived population estimates
For ~7,000 neighbourhoods in England & Wales, mean working age
population ~4,500
Monthly values 1985 - mid-2009 - so short-term effects visible Linked socio-economic characteristics of n'hoods (housing, occupation etc)
Research questions:
“Vulnerability” to recession Persistence and change Inequality
SLIDE 6
Neighbourhood claimant rates 1985-2009, selected percentiles Chart with deciles!
50 (second bottom line) is the median average rate 75 is the rate for the worst-off 25% 95 is the rate for the worst-off 5%
SLIDE 7 What does it tell us? JSA rates have fallen overall
Displacement onto incapacity benefits from early 1990s Problem for comparability
Small number of areas have very high rates Gap between worst-off and average grows in absolute terms in recessions
Long-term high-unemployment neighbourhoods suffer most More stark if considered as risk-per-person of becoming unemployed
SLIDE 8 Vulnerability to short-term shocks: Neighbourhood characteristics & rise in unemployment 2008-09
Long-term Unemployment Region
Base claimant rate London North East NS: North West Yorks & H East Mids West Mids East South East South West Wales
Industries & Workforce Neighbourhood
Manufacturing Finance Real Estate No qualifications No car Construction Public Sector Younger workers 16-30 Older workers 50+ High qualifications Social rented NS: Private rented Not White British
Orange bar = associated with bigger rise in JSA Green bar = associated with smaller rises
SLIDE 9 21 years and one-and-a-half recessions: long-term change? Use standardised rates for comparison over longer periods Simple correlation of 1986 rates with 2007 rates is .75 Of the 614 n'hoods which were in 1st (worst-off) decile in 1985:
400 were in the top decile again in
2005
Only 8 (1.9%) had below average
JSA rates in 2005
Starting point Region
Claimant rate Q2 1986 (NS) East Mids West Mids East South East London North East North West Yorks & Humb South West Wales
SLIDE 10 Cities and their high-unemployment neighbourhoods Some of the variation in rates is difference within cities and regions
About 75% of overall variance is that between better-off and worse-off
n'hoods within each town/city
About 25% (by one measure) is the variation between cities and regions Difference between cities was:
Highest in the late 1980s Lowest in the depths of the 1990s recession Has been gradually, though slightly, declining since 2000 Effects of the current recession not yet apparent
SLIDE 11
The trajectories of some small cities: Base neighbourhood JSA unemployment rate, relative to E&W
SLIDE 12 Project 2: Classifying high-unemployment neighbourhoods in 2008 Policy interest in use of spatial area classifications / typologies
Allocation of resources Selection of suitable interventions Use in evaluation - identifying similar 'control' areas
Statistical typologies have to be based on a selection of variables
But what is 'relevant' to concentrated unemployment depends on
perspective
Regional or local causes? Housing, migration or people? Etc
SLIDE 13 Project 2: Classifying high-unemployment neighbourhoods in 2008 Model three dimensions of employment deprivation at neighbourhood level, for worst 20% areas on IMD
Excess disability (IB/ESA claims) Claimant unemployment (JSA rates) Seasonal variation in unemployment (JSA flows)
Consider three spatial levels
Neighbourhood (LSOA): demographics, housing, labour force characteristics Housing market (LA): rents, migration, commuting Labour market (NUTS3): wages, productivity, labour demand
Use results of models as basis for cluster analysis
SLIDE 14 The JSA model The variance is both between and within labour markets
cf above: ~25% of variation is
between regions
Area and neighbourhood
characteristics both useful
Interactions: e.g. high rents + low
entry-level wages + social housing
What predicts JSA claim rates in most deprived 20% is not the same as in all n'hoods
Spatial level % Variance Labour Market (NUTS3) 24% Housing Market (LA) 14% Neighbourhood (LSOA) 62%
SLIDE 15
A four-way classification
Group Description
A
Highly deprived social housing neighbourhoods
B
Older workers in declining areas
C
High-churn neighbourhoods with younger workers
D
Ethnically mixed neighbourhoods in stronger labour markets
(E)
(Inner London)
SLIDE 16
A ten-way classification
Description 4 i
Social housing n'hoods with extreme multiple deprivation
A ii
Multiply deprived social housing n'hoods
A B iii
Dormitory, declining n'hoods in very weak economies; much ill-health
A B iv
Stable n'hoods with older workers, steady employment
B v
N'hoods with private housing in weaker self-contained labour markets
C B Description 4 vi
N'hoods with young population in vulnerable employment
C B vii
High turnover, socially mixed n'hoods in self-contained labour markets; much hospitality work
C viii
Mixed social housing n'hoods in buoyant cities
D ix
Young, socially and ethnically mixed n'hoods in buoyant cities
D x
Inner London
E
SLIDE 17 i Soc hsg, extreme depr ii Soc hsg, multiple depr iii Declining areas, older, IB iv Older wrkrs, stable emp v Weak self-cont markets vi Young pop, vuln work vii High turnover, soc mix viii Soc hsg mix in buoyant ix Soc / eth mix in buoyant x Inner London
The North of England
SLIDE 18 i Soc hsg, extreme depr ii Soc hsg, multiple depr iii Declining areas, older, IB iv Older wrkrs, stable emp v Weak self-cont markets vi Young pop, vuln work vii High turnover, soc mix viii Soc hsg mix in buoyant ix Soc / eth mix in buoyant x Inner London
The Midlands and the South East
SLIDE 19
The South West
SLIDE 20
High-unemployment neighbourhoods in smaller cities Percent of neighbourhoods by classification
England Hull Stoke Plymouth Bath 1 Extreme Deprived Soc Hsg 13 54 10 17 2 Multiple Deprived Soc Hsg 18 10 1 25 75 3 Declining, Ill Health 15 11 35 6 4 Older, Steady Work, Stable 17 27 23 5 Local Work, Private Hsg 11 6 22 2 6 Young, Vulnerable Work 7 6 4 10 25 7 Churn, Local Hospitality Work 5 13 15 8 Mixed Soc Hsg, Buoyant City 7 2 9 Young, Mixed, Buoyant City 5 * Inner London 2
SLIDE 21 Conclusions - neighbourhood unemployment High degree of unemployment persistence over ~25 years Cyclical unemployment effects strongly correlated with base unemployment
Reserve pools of labour, not cultures of worklessness?
Multiple spatial levels of analysis needed at once
~30-40% variance attributable to differences between labour markets Then - local demography / human capital / housing stock Interactions between neighbourhood housing tenure & wider area features Rented tenures, especially public housing, predominates
Mechanisms are different for highest unemployment areas
SLIDE 22 What does it mean for smaller cities? Very different trajectories Smaller cities have distinctive types of high-unemployment n'hoods
Varies by industrial history Varies by geographic features (self-containment) Varies by housing system (large estates? high-cost / low-cost
Implications for policy interventions
SLIDE 23 Further research Anthropological interpretation of poor neighbourhoods
E.g. somatic aspects of long-term withdrawal through ill-health Broader correlates - violence, teenage conception, educational motivation Neighbourhood sociology of unemployment
Prediction and forecasting
Recession effects, public sector cuts Benefits of regional development
Sociology of policy
The language of interventions Reducing, or managing, unemployment? What sort of regeneration is realistic in different cities?
