Monitoring Internal Migration in the United Kingdom

Monitoring Internal Migration in the United Kingdom

John Stillwell

School of Geography, University of Leeds

Presentation at the BSPS Annual Conference,

University of Leeds, 7-9 September 2015

Importance of internal migration

• Underpinned the processes of urban and rural change (urbanisation, rural exodus, suburbanisation, counterurbanisation, reurbanisation) that have defined the UK’s settlement pattern

• Critical component of population dynamics for local areas – magnitude and composition provide evidence of what services are required

• In recent years, tendency for attention to focus more on international migration

Presentation

1. Internal migration change in the UK 2001-11: based on data from the Census and annual estimates from administrative sources

2. Internal migration distance in England: based on a sample of data from a large commercial survey in mid-2000s

3. International comparison of internal migration: based on data from the IMAGE repository

Research questions

• How has migration changed in the UK over the decade between 2000-01 and 2010-11?

• What distance did internal migrants actually travel within England in the mid-2000s?

• How does the migration intensity and distance of migration compare with other countries?

• How important is scale in the analysis of migration indicators?

Key dimensions

Intensity – Migration rates Impact – Migration effectiveness (MEI) MEI = Net migration / (In-migration + Out-migration )*100

Distance – Mean migration distance Scale – Various

Key data sets

Census – 2001 and 2011 Administrative – PRDS/NHSCR/Estimates Survey – Acxiom Research Opinion Poll

Acknowledgements

• ESRC: UK Data Service-Census Support grant

• ESRC: PhD Case studentship with ONS for Nik Lomax – collaboration with Phil Rees and Paul Norman

• ESRC: PhD for Michael Thomas (supervised jointly with Myles Gould) under the TALISMAN programme and Acxiom Ltd for providing the data for this project

• ARC: IMAGE project and collaboration with researchers in Australia, Poland and China led by Martin Bell

• Kostas Daras, Nikola Sander, Carl Bradbrook, Claire Woodvine

1. Internal migration change in the UK between 2000-01 and 2010-11

District Ward Output Area 1991 11 tables (94 counts) 2 tables (12 counts)

2001 10 tables (996 counts) 5 tables (96 counts) 1 table (12 counts)

2011 83 tables (2,551 counts) 50 tables (2,171 counts) 50 tables (2,189 counts) Open: 5 tables (83 counts) Safeguarded: 39 tables (916 counts)

Safeguarded: 9 tables (353 counts)

Safeguarded: 3 tables (3 counts)

Secure: 39 tables (1,552 counts)

Secure: 41 tables (1,818 counts)

Secure: 47 tables (2,186 counts)

Evolution of Census Origin-Destination Migration data

How has migration changed in the UK over the decade between 2001 and 2011?

• Use set of 404 consistent areas:

Districts

England 324

Wales 22

Scotland 32

Northern Ireland 26

Levels of migration in the UK (those aged over 1)

2000-01 2010-11 Change %

Change

Inter-district 2,443,964 2,768,632 +324,668 13.3

Intra-district 3,627,034 4,046,769 +419,735 11.6

Internal 6,070,998 6,815,401 +744,403 12.3

But there were an additional 458,000 migrants with unknown usual residence at the beginning of the period in 2000 – these can be allocated proportionately

Immigrants 401,174 687,174 +286,000 71.3

Total 6,472,172 7,502,575 +1,030,403 15.9

Adjusted levels of migration in the UK 2000-01 and 2010-11 (Migrants aged over 1)

2000-01 2010-11 Change %

Change Inter-district 2,616,807 2,768,632 151,825 5.8

Intra-district 3,883,547 4,046,769 163,222 4.2

Internal 6,500,354 6,815,401 315,047 4.8

Immigrants 429,546 687,174 257,628 60.0

Total 6,929,900 7,502,575 572,675 8.3

Migration intensities (%) in the UK 2000-01 and 2010-11

2000-011 2001-11 Change % Change Inter-district 4.50 4.44 -0.06 -1.3 Intra-district 6.68 6.49 -0.19 -2.8

Internal 11.18 10.92 -0.26 -2.3

Immigrants 0.74 1.10 0.36 48.7

Total 11.92 12.03 0.10 0.9

Note: 2000-01 intensities adjusted for origin not stated

UK migration intensities by age group, 2000-01 and 2010-11

Note: 2000-01 intensities adjusted for origin not-stated

Changes in UK migration intensities between 2000-01 and 2010-11

Note: 2000-01 intensities adjusted for origin not-stated

Net migration by district, 2000-01 and 2010-11

Migration effectiveness by district, 2000-01 and 2010-11

MEI by district, ages 15-19, 2000-01 and 2010-11

MEI by district, ages 45-59, 2000-01 and 2010-11

MEI by district, ages 60-64, 2000-01 and 2010-11

MEI by district, ages 65-74, 2000-01 and 2010-11

Internal churn rates, 2000-01 and 2010-11

y = 1.1554x - 2.2889 R² = 0.8976

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0

Ch

urn

rat

e 2

01

1 (

pe

r 1

00

)

Churn rate 2001 (per 100)

y = 0.6096x + 6.4362 R² = 0.1718

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0

Imm

igra

tio

n r

ate

20

01

1 (

pe

r 1

00

0)

Immigration rate 2001 (per 1000)

Immigration rates, 2000-01 and 2010-11

Time series change

• Census is problematic: - those with origin unknown in 2001 - periodic nature • Use of administrative data from NHSCR/PRDS to build time

series for the UK at district scale from 2001-02 to 20110-11

Spreadsheet information system

Inter-district migration intensity during the 2000s

Changes in inter-district migration intensities between 2001-02 and 2010-11

Different aggregations of districts

e.g. City regions and their component parts: Metropolitan: - Core - Rest Non-metropolitan - Near - Coast & Country

Circular plots showing migration flows of over 280 migrants between and within city regions in 2001-02 and 2010-11

http://nikolasander.com/uk-migration/

2001-02 2010-11

Changes in migration between metro and non-metro aggregations of districts: time-series indices

Source: Stillwell, J., Lomax, N. and Sander, N. (2015) Monitoring and visualising sub-national migration trends in the UK, In Geertman et al. (eds) Planning Support Systems and Smart Cities, Springer

2. Internal migration distance in England, 2005-2007

Source: Stillwell, J. and Thomas M. (2015) How far do internal migrants in England really move?, Working Paper 15/1, School of Geography, University of Leeds

Acxiom ROP – voluntary lifestyle survey household respondents aged 18 and over

In 3 years (2005, 06 and 07), asked question about when did you move and where was previous residence

950,658 cases (movers and stayers) over 3 years Subset 1: 125,494 migrants Subset 2: 26,688 migrants in previous 12 months Subset 3: 43,390 migrants in previous 36 months

Key benefit: Postcodes of origin and destination available

Using Subset 1, how far do people move?

Using Subset 2, how does distance moved vary by migrant type?

Source: Thomas, M, Stillwell, J. and Gould, M. (2015) Modelling multilevel variations in distance moved between origins and destinations in England and Wales, Environment and Planning A, 47, 996-1014

Using Subset 3, how does postcode distance compare with centroid distance?

9 areas 27 areas 324 areas 6,793 areas

(i) Inter-area migration

(ii) Intra-area migration Intra-area distance calculated as the radius of a circle equivalent to the area divided by the square root of 2 (Batty, 1976)

How can we improve the measure of intra-area distance?

Use the relationship between inter-area migration distance and area size at different spatial scales?

Distance measure: Inter-centroid

distances plus:

2000-01 2010-11

𝑅2 IoD MAPD 𝑅2 IoD MAPD

(i) Conventional estimate for intra 0.908 27.2 55.7 0.921 25.6 51. 5

(ii) Linear estimate for intra 0.963 19.5 39.2 0.960 18.6 37.5

(iii) Quadratic estimate for intra 0.978 17.2 34.6 0.977 16.5 33.3

What are the results of fitting constrained spatial interaction models?

Inte

r-ar

ea m

igra

tio

n

dis

tan

ce (

km)

Linear model Quadratic model

Generate estimates of intra-area distance from area size using model equations which can be used in modelling 2001 and 2011 Census migration flows

3. International comparison of internal migration

IMAGE is an international collaborative program which aims to provide a robust basis for comparing internal migration between countries around the world

http://www.gpem.uq.edu.au/image

IMAGE Inventory Review of internal

migration data collection practice in the 193 UN member

states

IMAGE Repository Global collection of

population and internal migration data and GIS

boundaries for 135 countries

IMAGE Studio Analytical software to

compute internal migration measures

and address key methodological issues

RESEARCH FRAMEWORK

Major problem with cross-national comparison of internal migration

• Different shapes and sizes of the spatial units that are used for counting migration flows

• Modifiable Area Unit Problem (MAUP) which has identified two components (Openshaw, 1984):

- scale effect or the variation in results obtained when data for one set of areal units is aggregated into larger spatial units (i.e. where the number of regions changes)

- zonation or aggregation effect or the variation in results obtained from different ways of subdividing geographical space at the same scale (i.e. where the number of regions remains the same)

• IMAGE Studio is software that tries to address these effects

IMAGE Studio: Framework and Sub-systems

Aggregation sub-system Example of the IRA wave algorithm

This process can be repeated iteratively with different seed points

Example aggregation for the UK

0

500000

1000000

1500000

2000000

2500000

3000000

10

30

50

70

90

11

0

13

0

15

0

17

0

19

0

21

0

23

0

25

0

27

0

29

0

31

0

33

0

35

0

37

0

39

0

41

0

Me

an t

ota

l in

ter-

ASR

mig

ran

ts

Scale (Number of ASRs)

Start: N = 420 BSUs M= 2.5 million CMI = 4.2%

End: n = 10 ASRs M=1.1 million CMI = 1.9%

Why is the Studio useful? Example 1: Comparing Aggregate Crude Migration Intensities (ACMI) between countries

Cru

de

mig

rati

on

inte

nsi

ty

(CM

I)

Ln (Households/Number of ASRs)

In some countries (Country A), CMI maybe available from data at different spatial scales, including total migration which gives ACMI

Country A Linear relationship is known as Courgeau’s k

In other countries (Country B), total migration data are not available but CMI may be available at one scale (e.g. districts)

Studio can be used to generate CMI at different spatial scales and regression line used to identify ACMI

Country B

One-year migration league table of aggregate intensities UK

Source: Bell, M., Charle-Edwards, E., Kupiszewska, D., Kupiszewski, M., Stillwell, J. and Zhu, Y. (2015) Internal migration and development: comparing migration intensities around the world, Population and Development Review, 41(1): 33-58

Example 2: Comparing migration distance and distance decay between countries

0 400000 800000

Mean ASR population

Mean migration distance (MMD)

Mea

n b

eta

0

1

2

3

Beta derived from model is as follows:

𝑀𝑖𝑗 = 𝐴𝑖 𝐵𝑗 𝑂𝑖 𝐷𝑗 𝑑𝑖𝑗−𝑏𝑒𝑡𝑎

4 6 8 10 12 14

Ln (Mean ASR area)

Mea

n M

MD

0

50

0

1

00

0

15

00

Distance decay parameter (Beta)

Example 3: Comparing migration distance and effectiveness in the UK by age group, 2001 and 2011

70

90

110

130

150

170

190

210

230

10

50

90

13

0

17

0

21

0

25

0

29

0

33

0

37

0

40

4

2001

2011

50

100

150

200

250

300

10 40 70 100130160190220250280310340370400

2001

1 to 14

15 to 19

20 to 24

24 to 34

35 to 44

45 to 59

60 and up

50

100

150

200

250

300

10 40 70 100130160190220250280310340370400

2011

1 to 14

15 to 19

20 to 24

24 to 34

35 to 44

45 to 59

60 and up

Distance

Number of ASRs Number of ASRs

Number of ASRs

Example 3: Comparing migration distance and effectiveness in the UK by age group, 2001 and 2011

National MEI = 100 ∑|Net migration| / ∑ (In-migration + Out-migration)

Effectiveness 2001 2011

Conclusions • Census and administrative data suggest modest declines

in internal migration intensities in the UK over the 2000s, most marked for those in the latter half of the age range (but impact of recession not that significant)

• Changes in net migration suggest spatial pattern of waning counterurbanisation and increasing re-urbanisation

• Inter-district migration distance decreasing

• Convergence of migration impact by age group at lower level, but MEI increasing for 15-19 year olds (students)

• Compared with other countries for which one-year data are available, UK has average aggregate intensity

Take-away messages

Migration researchers should:

• make greater use the under-exploited origin-destination migration data sets that have become available from the 2011 Census

• investigate the potential of less conventional data sources for migration analysis particularly in the event of a possible shift away from census-taking after 2021

• attempt more cross-national comparative research on internal migration (perhaps using data collected as part of the IMAGE project)

• give more consideration to issues of scale and zonation when doing migration research