Transcript Tony: here are `chunks` of poss. use

Paper presented at the BSPS Annual Conference, Leicester,
13-15 September 2004
Population change 1971-2001 for
the City Regions and Localities of
Great Britain
Tony Champion and Mike Coombes
Centre for Urban & Regional Development Studies
University of Newcastle, Newcastle upon Tyne, NE1 7RU
Tel: +44 (0) 191 222 6437, Email: [email protected]
Population change 1971-2001 for the City
Regions and Localities of Great Britain
• Introduction: aims, outline, acknowledgments to
Simon Raybould & Colin Wymer
• The City Regions & Localities framework
• Population change 1971-2001 for City Regions:
growth, decline, recovery, backsliding?
• City Region cores as leaders or laggards in City
Region growth?
• Between-Locality variations in change rates:
testing roles of North/South & Urban/Rural
• Summary and next steps in research
The City Regions & Localities framework
• Derived from research for ESRC by CURDS:
see Coombes (2000)
• Defined on the basis of information on
functional linkages and areal associations
• 307 Localities which typically comprise at least
one urban centre and adjacent linked areas
• 43 City Regions centred on Localities with
‘regional city’ characteristics and containing
others linked by commuting and migration
• Embraces polycentric urban forms
Rationale behind this framework
• Administrative areas are the ‘off the shelf’ option but
they thwart meaningful city/town comparisons:
! Manchester is under- and Leeds over-bounded !
? which part of London is comparable to Bristol ?
? what can be used as a set of City Regions ?
• Research shows coherent patterns emerge in analyses
if meaningful functionally-defined areas are used
* areas reflect how modern urban systems work
* each area can be seen as a local housing market
* the local ‘churn’ of house moves is internalised
* hierarchy of Localities fitting into City Regions
Defining functional areas in practice
• Need to analyse data on patterns of flows / linkages:
commuting migration goods information etc
• Definitions of ‘multi-function’ areas should ideally
reflect the patterns in several varied flow datasets
• The ‘traditional’ model was of one ‘core’ with its
catchment area
but polynuclear areas are now more common,
so analyses must look at flows to / from everywhere
• The 307 Localities are defined using a Synthetic
Data method drawing on very many different
strands of evidence on area linkages
• Example: in and around W Yorks ‘urban area’
West Yorkshire: complex case study
Localities: clustered linkages = ‘towns’?
City Regions: groups of Localities
City Regions in northern England
City Regions in southeastern England
38 City Regions of England and Wales
Population change 1971-2001
for City Regions
• Data from Census, using (mainly) EDs/OAs to
produce best-fit to the 1991 wards from which
Localities and City Regions are built
• Population defined on the basis used at each
Census, i.e. population present 1971, residents
(present/absent basis) 1981, residents 1991,
residents (students at term-time address) 2001
• Population change rate adjusted to % point
above/below GB rate (to measure change in
relative performance between decades and
partially allow for differences in bases)
How similar are City Region 1991-2001 change rates
from Census and MYE (latter for best-fit from LAs)?
1991-2001 population change (%) for 43 City Regions,
census-based versus mye-based
8
Cambridge
ymye-based
change
6
4
Oxford
2
York
0
-8
-6
-4
Inverness
-2
0
2
-2
-4
-6
-8
-10
census-based change
4
6
8
10
How similar are City Region 1981-1991 change rates
from Census and MYE (latter for best-fit from LAs)?
1981-91 population change (%) for 43 City Regions,
census-based versus mye-based
15
mye-based change
10
5
Aberdeen
Inverness
0
-10
-5
0
-5
-10
census-based change
5
10
15
Population change for City Regions, 1971-2001,
standardised to GB rate = 0.0%: top and bottom five
Rank
1971-1981
% 1981-1991
% 1991-2001
%
1
Peterborough
19.6 Peterborough
12.1 Cambridge
7.3
2
Northampton
12.7 Cambridge
9.5 Oxford
6.5
3
Shrewsbury
12.0 Exeter
7.8 Worcester
5.3
4
Norwich
10.8 Oxford
7.7 Shrewsbury
5.0
5
Cambridge
10.6 Northampton
7.7 Exeter
5.0
(GB)
(0.0) (GB)
(0.0) (GB)
(0.0)
39
Manchester
-3.2 Sheffield
-4.0 Manchester
-3.8
40
Dundee
-3.2 Manchester
-4.1 Newcastle
-4.8
41
London
-3.3 Newcastle
-4.6 Middlesbrough
-5.0
42
Liverpool
-5.0 Glasgow
-7.6 Glasgow
-5.8
43
Glasgow
-6.9 Liverpool
-8.0 Liverpool
-7.0
43 City Regions: 1981-1991 and 1991-2001
standardised change rates compared
population change rate, 43 GB City regions, 1981-1991 versus 1991-2001,
% point differentials from GB rate
8
1991-2001
4
0
-10
-5
0
5
-4
-8
1981-1991
10
15
% point shift in population change for City Regions,
1970s-to-1980s and 1980s-to-1990s,
standardised to GB rate = 0.0%: top and bottom five
Rank
1970s to 1980s
% pt
1980s to 1990s
% pt
1
London
3.71
Leeds
2.63
2
Oxford
2.81
Coventry
2.54
3
Brighton
2.78
London
2.22
4
Exeter
2.59
Lincoln
1.99
5
Southampton
1.95
Edinburgh
1.90
39
Middlesbrough
-4.38
Norwich
-3.96
40
Norwich
-4.75
Plymouth
-5.88
41
Shrewsbury
-4.78
Aberdeen
-6.25
42
Northampton
-4.97
Peterborough
-7.76
43
Peterborough
-7.49
Inverness
-8.33
Classification of 43 City Regions by level of
change rate for 1970s, 1980s and 1990s
• Steady growth (rate above GB in all decades);
Brighton, Bristol, Cambridge, Derby, Exeter, Gloucester,
Ipswich, Leicester, Lincoln, Northampton, Norwich,
Oxford, Peterborough, Plymouth, Portsmouth, Reading,
Shrewsbury, Southampton, Worcester, York (20)
• Steady decline (rate below GB in all decades):
Birmingham, Bradford, Cardiff, Coventry, Dundee,
Glasgow, Leeds, Liverpool, Manchester, Newcastle,
Sheffield (11)
• Recovering (rate shifts from below to above GB):
Edinburgh, London (2)
• Backsliding (rate shifts from above to below GB):
Aberdeen, Carlisle, Chester, Hull, Inverness, Middlesborough, Nottingham, Preston, Stoke, Swansea (10)
• NB. No moves across GB=0.0 line and back again
Classification of 43 City Regions by trajectory of
standardised change rate 70s-to-80s, 80s-to-90s
• Upward shift between both pairs of decades (//):
Edinburgh, London (2)
• Upward shift into 1990s after downward in 1980s, i.e.
troughing in 1980s (\/):
Bradford, Coventry, Derby, Glasgow, Leeds, Leicester,
Lincoln, Liverpool, Manchester, Sheffield (10)
• Downward shift in 1990s after upward in 1980s, i.e.
peaking in 1980s (/\):
Brighton, Dundee, Exeter, Oxford, Plymouth, Portsmouth,
Southampton, Worcester, York (9)
• Downward shift between both pairs of decades (\\):
All others (22)
City Region ‘cores’ as leaders or laggards
in City Region growth?
• Concerning the current debates about investing in
‘core cities’ so as to revive the regions
• [Alternative academic debate: are City Regions
(de)concentrating in absolute or relative terms?]
• To what extent are City Region cores growing more
strongly than (the rest of) their Regions?
• How has the performance of the cores relative to their
City Regions altered since the 70s?
• Definitions: ‘core’ = Regional City of each City Region;
performance = population change (from Census, with
checks against Mid Year Estimates, MYEs)
Population change differential between Regional City and
its City Region, 1971-81, 1981-91 and 1991-2001
Population change differential between Regional City and its City Region, 1971-81, 1981-91 and 1991-2001
10
% point difference (RC minus CR)
RC-CR7181W
RC-CR8191E
RC-CR9101E
5
0
-5
-10
City region ranked by RC minus CR 1991-2001 rate differential
City Regions classified by whether
Regional City was lagging or leading
Size of
CR in
2001
(million
people)
Number
of CRs
Regional City lagging
Regional City leading
19711981
19811991
19912001
19711981
19811991
19912001
All CRs
43
35
35
21
8
8
22
3.20+
1
1
1
0
0
0
1
1.60-3.19
7
7
7
3
0
0
4
1.00-1.59
8
6
5
1
2
3
7
<1.00
27
21
22
17
6
5
10
Population change, 1971-2001, for ten major cities and their City Region
remainders
15.0
Core 71-81
Rest 71-81
Core 81-91
Rest 81-91
Core 91-01
Rest 91-01
10.0
% for decade*
5.0
.0
London
Birmingham Manchester
Liverpool
Bristol
Leeds
Sheffield
Glasgow
Nottingham
-5.0
-10.0
-15.0
-20.0
*NB. - actual rates (GB rate 1971-81 -0.7%, 1981-91 2.5%, 1991-2001 4.0%)
Newcastle
Population change, 1981-2001, for ten major cities and their City Region
remainders (calculated from MYEs 4.11.03 using best-fit LAs)
10.0
Core 81-91 MYE
Rest 81-91 MYE
Core 91-01 MYE
Liverpool
Leeds
Rest 91-01 MYE
8.0
6.0
% for decade*
4.0
2.0
.0
London
Birmingham Manchester
Bristol
Sheffield
Glasgow
Nottingham
-2.0
-4.0
-6.0
-8.0
-10.0
*NB. - actual rates (GB rate 1981-91 1.9%, 1991-2001 2.7%)
Newcastle
City Region ‘cores’ as leaders or laggards
in City Region growth? Main results:
• For 43 City Regions, Regional Cities were predominantly
laggards in 1970s and 1980s, but half were leaders by
1990s
• Among CRs with over 1m residents in 2001, the majority
(12 out of 16) had leading RCs in 1990s
• Among smaller CRs, the majority (10 out of 27) had
lagging RCs in 1990s
• Among GB’s 10 major cities, all had lagging RCs in 1970s
and 1980s, but in 1990s 6 had leading RCs (Leeds,
Liverpool, London, Newcastle, Nottingham, Sheffield)
BUT (allowing for 1991 undercount and ‘student’ effect)
• Using MYEs for LA-best-fits, only London CR had a
leading CR in 1990s
Between-Locality variations in change rates:
testing the roles of North/South & Urban/Rural
• Traditionally, the principal dimensions are North v
South, Urban v Rural – how has the strength of
these altered since 1970s?
• North/South split based on line roughly from
Severn to Humber: South includes CRs of
Coventry, Leicester, Nottingham and Lincoln
• Urban/Rural based on Coombes & Raybould’s
Urbanization Index 1991
• Coverage is 262 Localities of England and Wales
• Graphs, then regression results for N/S & U/R
• Analysis by previous-decade rates and N/S
% population change for 262 Localities of
England and Wales, by log Urbanization
Index 1991 and North (red) / South (green)
1981-1991
1991-2001
30
40
30
20
20
10
10
0
%91-01E
0
-10
-.5
0.0
log(UINDX91)
.5
1.0
1.5
2.0
NS
NS
2
2
1 -10
1
-.5
0.0
log(UINDX91)
.5
1.0
1.5
2.0
262 Localities of England & Wales:
regression results for N/S & U/R
Decade
Beta coefficients
Adjusted
R2
N (1) / S (2)
Log UrbIndex91
1971-1981
0.283
-0.230
0.137
1981-1991
0.370
-0.546
0.464
1991-2001
0.449
-0.341
0.338
 Positive effect of South rises between decades
 Negative effect of Urban much stronger in 80s than 70s, but
drops back in 90s
 Low R2, so plenty of variance to be accounted for by other
factors
262 Localities of England & Wales:
regression results for N/S & momentum of
previous-decade population change rate
Decade
Beta coefficients
Adjusted
R2
N (1) / S (2)
Rate for previous
decade
1981-1991
0.243
0.567
0.460
1991-2001
0.266
0.509
0.438
 Considerable momentum from previous decade, though
slightly less from 80s to 90s (student effect?)
 Simple correlation of change rates between decades:
80sV70s r=0.64, 90sV80s r=0.62
 Positive effect of South rises a little between decades
 Still a lot of variance in between-decade shift to be
‘explained’ using other factors, more in 90s than 80s
Summary of findings
• Applied a new portrayal of GB’s urban system that
chimes with current policy initiatives on regional
development and housing market areas, as well as
academic research
• Confirms persistent underperformance of City
Regions headed by large provincial cities, while
strongest are CRs in much-greater-SE and Midlands
• Across 3 decades London and Edinburgh are in a
class of their own: decline->growth, accelerating
• Regional Cities shifting to lead their CRs in 90s,
according to Census data, but not such an
encouraging picture for 10 major cities from MYEs
• At Locality level, clear N/S and U/R dimensions to
growth in all 3 decades, with N/S increasing in
importance but U/R peaking in 80s – and less inertia
in patterns 80s-to-90s
Next steps in research
• Further tests on robustness of change rates,
using latest (09.09.04) MYEs … but challenges of
adjusting for Census undercounts in both 1991
and 2001 as well as student definition change,
and of allowing for LA v ED/OA effect
• Further work on trends across 3 decades in
RC/CR lead/lag relationships and discriminating
factors in relative performance of RCs
• Annual trends, using MYEs (after 07.10.04)
• Analysis of the ‘unexplained’ variance in growthrate variance at Locality level
• Examination of alternative performance measures
such as job growth and GVA
• Comparison of insights with those from other
geographies, e.g. ‘urban areas’ from SOCR 2005