Ethnic change and diversity in England, 19812001

1
Ethnic change and diversity in England, 1981-2001

• Phil Rees and Faisal Butt
• University of Leeds
• RGS-IBG Annual Conference
• London 3-5th September 2003

2
Introduction

• The paper develops an ethnic population database
  covering the three censuses of 1981, 1991 and
  2001 using common group and area definitions. The
  database makes possible a comparison across two
  decades.
• Englands population is changing its ethnic
  composition rapidily. The 2001 Census of
  Population enables us to examine this change for
  the first time with any certainty at spatial
  scales smaller than region level (Rees 2001
  Scott, Pearce and Goldblatt 2001).

3
Aim

• The aim to examine the first and second
  derivatives of ethnic change.
  
• That is, to look at the spatial structure of
  ethnic population changes over two inter-censal
  decades and to interpret those changes in a
  provisional way.

4
Key geographical questions needing answers

• Having constructed this roughly comparable, three
  census
  
• data set, we use it to address a number of
  interesting
  
• questions.
• How has the ethnic composition of the England
  population changed? How have the ethnic
  compositions of the different regions changed?
  How diverse is the population in the different
  regions and how is diversity changing?
• Are the different ethnic groups living more
  closely together or further apart? In particular,
  is there spatial separation between the white
  majority and ethnic minorities? Have these
  separations been increasing or decreasing?

5
Key geographical questions needing answers
continued

• Looking at each of the ethnic groups, to what
  degree is each group concentrated in the largest
  cities? What spatial shifts have been occurring?
  
• To what extent is each ethnic group concentrated
  in London? Is London unique in terms of its
  ethnic composition? How has this been changing?
  To what extent is each group participating in
  outward shifts within the capital region?
• What is the relationship between spatial
  concentration of a group and change in the past
  decade? Is growth occurring in areas of highest
  concentration or in areas of lowest?

6
Data and estimation methods

• The database of ethnic populations constructed in
  this paper spans three censuses.
  
• There were two tasks that had to be accomplished
  the creation of common ethnic groups in 1981,
  1991 and 2001 and the creation of common spatial
  units at each census.

7
How common ethnic groups were defined

• For the 1981 Census ethnic populations we rely on
  estimates for local authorities in Great Britain
  made by Rees and Phillips (1996, Appendix 2.2,
  pp.85-109).
• The 1991 ethnic group estimates were adjusted to
  the 1991 mid-year estimates revised in the light
  of the 2001 Census for the total population,
  having adjusted for under-enumeration
  differentially by ethnic group using 1991
  ratios.
• In the 2001 Census the principal ethnic group
  classification comprises sixteen groups (see
  Table 1). The mixed group was split back into
  its constituent parentages and the Other group
  assigned in the ways set out in Box 1.

8
Box 1 Conversion of the 2001 Census ethnic
categories into 1991 Census ethnic categories
9
How common geographical areas were defined

• Common spatial units from the 2001 census were
  adopted for 1991 and 1981 using a rough method of
  geographical conversion.
  
• Counts from the 1991 Census of four aggregated
  groups (White, Black, Asian and Other) were
  extracted from the Linking Censuses through Time
  (LCT) website on a 2001 Census Local and Unitary
  Authority geographical base.
• Using information from the Office for National
  Statistics (ONS 1999, 2003), Local Authorities
  existing at the time of the 1991 Census were
  assigned on a best fit basis to 2001 Census Local
  and Unitary Authorities.
• The final step in the estimation was to adjust
  the ethnic group populations in 1991 for Local
  Authorities so that they summed to the revised
  estimates published by ONS for 2001 defined Local
  Authority units (ONS 2003). The steps in the
  estimation are set out as formal equations in Box
  2.

10
(No Transcript)
11
How common geographical areas were defined
continued..

• Clearly, with the volume of adjustments
  incorporated in the generation of the
  1981-1991-2001 harmonized ethnic population
  database and the large number of implied
  assumptions, the resulting estimates are going to
  be approximate.
• Note that because the analysis is mainly carried
  out at metropolitan region/region remainder
  scale, the effect of errors at Local and Unitary
  Authority level will be dampened.
• However, we argue that although there is some
  resulting fuzziness in the indicators used, the
  patterns revealed of ethnic change and
  redistribution are so clear and plausible, we can
  have a high level of confidence in the estimates.

12
The national (England) picture
Englands ethnic composition more varied in 2001
than ever before.

• Before looking at the spatial redistribution
  using the harmonised but fuzzy database, let us
  examine the make-up of Englands population in
  terms of 2001 Census ethnic group.

13
A myriad of national groups now living in England
populations in 2001

• These are the groups that can be easily
  identified but the composite categories are also
  important
  
• Just how diverse are the origins of Englands
  populations is made clear from Table 2s
  statistics.
  
• Each of these country of birth groups has its own
  migration history and spatial distribution in the
  UK which the detailed univariate tables of the
  2001 Census will enable researchers to study at
  fine spatial scales

14
(A) Changes in Ethnic Shares
15

• (B) BEM and White changes
• The first point to make is that the White
  population is hardly growing at all only by
  0.4 in the 1981-1991 decade and by 0.2 in the
  1991-2001 decade.
• Overall, the Black and Ethnic Minority (BEM)
  population maintained the pace of growth between
  the two decades 41 between 1981 and 1991, 39
  between 1991 and 2001. This average, however,
  conceals groups experiencing acceleration (Black
  African, Pakistani, Chinese) and groups
  experiencing deceleration (Black Caribbean and
  Black Other combined, Indian, Bangladeshi, Other
  Asian and Other groups).
• The White share of the population falls from
  95.4 to 91.4 over the twenty years while the
  BEM share rises from 4.6 to 8.6. The relative
  shift is accelerating with a -1.7 fall for the
  White share in 1981-1991, compared with a -2.3
  fall in 1991-2001. The Diversity Index (explained
  later in connection with Table 8) nearly doubled
  from 0.08 to 0.16.

16
(C) Individual Groups

• The Black population is, by contrast, growing
  rapidly and this growth has accelerated from 30
  in 1981-1991 to 40 in 1991-2001. There is a
  strong immigration flow and very high fertility
  in this group.
• The South-Asian origin population has doubled
  over the 1981-2001 period, growing slightly
  faster than the Black ethnic group as a whole but
  not as fast as the Black African group.
• Although the Indian group is the largest ethnic
  minority, its growth has slowed from 38 in
  1981-1991 to 27 in 1991-2001 (as a result of
  lower net immigration, lower fertility and higher
  mortality because of ageing).

17

• (C) Individual Groups continued
• The Bangladeshi population has been growing very
  rapidly (high fertility, young fertile age
  structure, continued net immigration) but this
  growth is decreasing (76 growth in 1991-2001
  compared with 96 in 1981-1991).
• The Pakistani populations growth, although
  significantly lower than the Bangladeshi group,
  has risen a little in the 1991-2001 decade, 56
  compared with 43 in the earlier decade.
• The Chinese population has doubled in the twenty
  year period 1981-2001 with higher growth 51 in
  the second decade than in the first (41). Higher
  immigration is probably the driver here.

18
The regional picture

• At the start of the paper we asked whether
  Englands ethnic groups
  
• were polarising that is, becoming more
  spatially separated.
  
• Between 1991 and 2001 Whites experienced an
  increase in an index of dissimilarity with the
  rest of the population while most ethnic minority
  groups experience decreased (more mixing amongst
  the BEM groups) (Dorling and Rees, 2003).
• Here we extend that analysis by focussing on the
  relationship between Whites and each BEM group.
  This is probably the most important inter-group
  relationship, between the host population and the
  immigrant origin groups.
• The Indexes of Dissimilarity are calculated. In
  this case the indexes are computed using the
  twenty metropolitan/remainder regions.

19

• Whites and BEMs further apart but BEMs spreading
  out
  
• The 1981-1991 decade saw a universal increase in
  the spatial polarisation of Whites and BEM
  members, ranging from an increase of 2.1 in the
  index of dissimilarity for the Black Caribbean
  group to 7.8 for the Other Asian group.
• In the 1991-2001 decade, either the indexes
  decreased (Black Caribbean, Black African,
  Indian, Chinese, Other Asian) or presented lower
  increases (Pakistani, Bangladeshi, Other Groups,
  the BEM as a whole).
• We can offer the following provisional
  interpretation of the changes between decades. In
  the 1980s many BEM groups were engaged in family
  building (more people in the same localities) or
  consolidation of pioneer locations through
  further immigration.
• In the 1990s new household formation and
  socioeconomic progress for BEM groups led to
  expansion in areas outside of the pioneer
  locations and a decrease in spatial polarization
  compared with Whites.

20
Metropolitan concentration now moving into
deconcentration

• Virtually all studies of the BEM groups have
  commented on the urban concentration of
  immigrant-origin groups.
  
• We divide the twenty regions of England into two
  groups metropolitan regions and Government
  Office Region remainders.
  
• For the BEM groups the 1981-1991 decade was
  characterised by gains in the share of the
  population living in metro areas against a
  background of very substantial growth in absolute
  numbers overall.
• The 1991-2001 decade shows a dramatic turnaround.
  All BEM groups (except the Black Other and Other
  Asian groups) exhibit declining shares of their
  populations resident in metropolitan regions.
  Again, this is against a background of continued
  substantial growth
• The British population has been experiencing
  deconcentration from its largest cities since the
  1960s and in the 1990s established BEM groups
  joined this trend. The relative growth of the
  BEM population was faster outside metro areas
  than within them.

21
London the most important city for most ethnic
groups

• There is considerable variation across the
  minority ethnic groups in their concentration in
  London and its constituent parts. The Pakistani
  group has a low level of concentration in London
  with a share comparable to Whites and the total
  population, but its London share increases over
  the twenty years.
• At the other extreme the Black groups have 60 or
  more of their population concentrated in London
  with four out of five Black Africans in England
  living in London. The concentration, however,
  decreased in the 1991-2001 decade for London as a
  whole and for Outer London.
• London thus plays a vital role in creating a new
  multi-ethnic and multi-cultural England.
  Reviewing the remaking of the Londoner 1948-99
  White (2002, p.167) concludes
• Forty years on, the most dramatic changes in
  the history of the Londoner have worked
  themselves through with patterns of collective
  give and take that remained recognizable from the
  earlier years of the century. In the process,
  they created a culture hugely enriched by a world
  of diversity.

22
Growing Diversity

• Diversity is variety. If a population is
  ethnically diverse, then it has both a large
  number of different groups and each of these
  groups has a non-negligible share of the total
  population.
• There is a substantial literature on the
  measurement of species diversity in biology. One
  of the most commonly used measures is the
  normalized entropy index (see Hunter College 2003
  for a discussion). However, here we employ a
  slightly simpler measure, the Diversity Index.
  For the twenty regions and ten ethnic groups and
  three years used in our harmonised database the
  correlation between the Entropy and Diversity
  indexes is 0.99.
• Diversity increased in all regions in the
  1991-2001 decade and in all but the last two
  remainder regions in the 1981-1991 decade. For
  England as a whole Diversity increased from 0.089
  in 1981 to 0.122 in 1991 (37 up) and to 0.163 in
  2001 (34 up).

23
The local picture

• The regional picture disguises, of course, a very
  wide variation within each region between local
  authorities and within local authorities between
  neighbourhoods.
• We select three ethnic groups with contrasting
  spatial distributions and use one index, the
  location quotient, to capture the 2001 situation
  and one index, the ratio of the 2001 population
  to the 1991 expressed as a percentage, to capture
  change.
• We use a cartographic device developed by Dorling
  et al. (2003). This consists of an EXCEL
  worksheet, the cells of which are used to
  represent statistical areas with equivalent
  populations.

24
Black Africans London centred but growing
everywhere

• The Black African group is almost exclusively
  located in London and a few Outer Metropolitan
  Areas.
  
• Wherever in the country it is located, the Black
  African group is expanding, probably through
  migration of students, family members and through
  new children (the group has a very high fertility
  rate).

25
Pakistanis located in northern and midlands
conurbations, but starting to move out

• The map of Pakistani location quotients is very
  different from that of Black Africans. Group
  members are concentrated in textile towns and
  cities of East Lancashire and West Yorkshire and
  in the engineering centres of South Yorkshire and
  the West Midlands.
• With industrial restructuring and decline in
  these staple industries the Pakistanis of working
  age have suffered high unemployment and a need to
  seek new opportunities.

26
Chinese the most dispersed ethnic minority group

• The Chinese group is concentrated in the London
  region but at only half the rate of Black
  Africans. In the West Midlands the LA with the
  highest concentration is not Birmingham but
  rather Solihull . This suggests that group
  members have made substantial progress up the
  socioeconomic ladder.
• This interpretation is supported by the map of
  change where the highest growth is happening in
  the belt of prosperous and attractive LAs and UAs
  that fill in the gaps between the major
  metropolitan centres.

27
Discussion

• Five summary points about 1981-1991 change were
  made in the earlier
  
• account. Here we assess whether the 1991-2001
  decade saw more of
  
• the same or evidence of a change in
  direction.
  
• (i) There is a profound contrast between the
  almost static White population , which grew by
  less than 1 percent between 1981 and 1991 and the
  minority ethnic populations which grew between 24
  and 95 percent.
• The 1991-2001 decade saw the same broad changes
  continue. It will be some time before the
  demographic momentum generated by youthful age
  structures and continued net immigration slows
  down as a result of ageing. But a proper
  demographic projection is needed to trace the
  future path of change properly.
• (ii) the degree of spatial redistribution
  was relatively modest. The geographical spread of
  each minority ethnic group in 1991 was
  substantially the same as it had been in 1981.
• Redistribution in the 1991-2001 decade was
  relatively modest again but there were
  significant shifts in direction. For most ethnic
  minority groups the share of their population
  residing in metropolitan regions began to
  decline, whereas in 1991 there was a greater
  degree of metropolitan concentration than in 1981.

28

• (iii) All groups were marginally more
  separated from Whites in 1991 compared with 1981,
  across a set of 20 metropolitan and
  non-metropolitan areas.
• The 1991-2001 saw a breakdown in this uniform
  trend. Although the BEM population as a whole was
  more segregated from Whites in 2001 than in 1981,
  the change was small. In five of ten groups the
  separation from Whites diminished.
• (iv) This increased concentration can be seen as
  a product of several processes.
  
• The five processes mentioned were White
  out-migration from areas of high BEM
  concentration, family growth, new household
  formation, new arrivals joining existing
  communities and relocation from economically
  declining areas to growing areas. These
  processes continued in the 1991-2001 decade.
  However, there was evidence of the start of
  spatial dispersion of many minority ethnic
  groups. Successful BEM households are joining the
  general population pattern of suburbanization and
  metropolitan deconcentration.
• (v) There is evidence of modest decentralization
  within metropolitan areas of ethnic
  concentration.
  
• The modest signs have swelled in the 1991-2001
  decade to more significant shifts. Six of ten
  BEM groups were shifting into the Outer
  Metropolitan Area faster than Whites.

29
Conclusions

• The rate of change has remained high and stable
  between 1981-1991 and 1991-2001.
  
• Some groups have grown rapidly (Black Africans,
  Bangladeshis), while others have seen moderate
  expansion (Indians, Other Asians).
  
• The White population has hardly grown and the
  White British population has probably declined.
  
• Black and Ethnic Minority (BEM) populations
  remain concentrated in metropolitan areas and
  under-represented in non-metro regions in 2001 as
  they were in 1991 and 1981. Whereas in 1981-1991
  BEM groups were concentrating into metropolitan
  areas, in 1991-2001 deconcentration began for
  most groups.

30
Conclusion continued..

• Within the metropolitan areas the London region
  stands out as highly dominant, housing more than
  50 of BEM populations as a whole in 1991 and
  2001, with at least 40 of individual minority
  group population residing in London with one
  exception, that of Pakistanis, who are still
  mainly located in the textile and engineering
  cities of the North and West Midlands.
• However, in 1991-2001 there is clear evidence of
  widespread deconcentration of BEM groups from
  their core areas, even if numbers are as yet
  small. Over the two decades the consequence of
  BEM population growth and spread in the second
  decade was a dramatic increase in ethnic
  diversity in all regions.
• The report on the 1981-1991 changes failed to
  monitor diversity. The current analysis shows
  that ethnic diversity continues to increase in
  most places at a steady pace. Every time we go
  out to dine in one of Englands cities we can be
  very grateful for this.

31
Supplemental Slides
32
(No Transcript)
33
(No Transcript)
34
(No Transcript)
35
Figure 1 Location quotients for ethnic groups,
1981-1991-2001, England White and Black Caribbean
36
Figure 2 Location quotients for ethnic groups,
1981-1991-2001, England Black African and Black
Other
37
Figure 3 Location quotients for ethnic groups,
1981-1991-2001, England Indian and Pakistani
38
Figure 4 Location quotients for ethnic groups,
1981-1991-2001, England Bangladeshi and Chinese
39
Figure 5 Location quotients for ethnic groups,
1981-1991-2001, England Other Asian Other Groups
40
(No Transcript)