Population Estimates for England and Wales BSPS Seminar Population Estimates Unit

1
Population Estimates for England and WalesBSPS
SeminarPopulation Estimates UnitImproving
Migration and Population Statistics
(IMPS)www.statistics.gov.uk/popestwww.statisti
cs.gov.uk/imps
2
Aims of the Seminar

• To gain a better understanding of the population
  estimates methodology
• To be aware of the sources used
• To allow delegates to contribute ideas and
  expertise on two important issues relating to
  population estimates Quality Assurance and
  Defining Usual Residence.
• To provide an overview of the work being
  undertaken as part of the Improving Migration and
  Population Statistics (IMPS) project.

3
Population Estimates - Overview

• The ONS population estimates
• National level for England, Wales
• Below national level, regions and local areas
• including local authorities (LAs) and primary
  care organisations (PCOs)
• By age and sex
• By marital status
• For total population and for household population
• Annual one year in arrears (mid-2005 due to be
  published 24 August 2006)

4
Population Estimates - Definitions

• The ONS population estimates are of the resident
  population and are as at 30 June
• The national resident population is
• all those usually resident regardless of
  nationality
• members of HM and US armed forces stationed in
  England and Wales. HM forces stationed overseas
  are not included.
• Currently looking at new definitions e.g. daytime
  population, weekday population.

5
Population Estimates Uses

• The key uses of the population estimates
• Base for the population projections that are used
  for local government resource allocation
• Also used as denominators in resource allocation
• Planning by central and local government e.g.
  services for the elderly
• Monitoring - used mainly as denominators in rates
  and ratios e.g. fertility and mortality rates
• Grossing up survey results e.g. Labour Force
  Survey
• Research by academics and others
• Commercial Usage

6
Population Estimates Methods 1

• Cohort component method Pt Pt-1 Bt - Dt
  It
• Population Base (aged on) Births - Deaths
  net Migration
• Most recent Census as the base (adjusted),
• Then update annually by adding births,
  subtracting deaths and allowing for net migration
• National and subnational, top down.

7
Population Estimates - Methods 2 Static
Populations

• These are sub-populations that dont age-on and
  are not covered by our estimates of Internal
  and/or international migration
• Armed Forces, Prisoners, School Boarders
• Method
• Remove these sub-populations (previous year)
• Age on rest of population
• Add in new sub-populations (current year)

8
Census as a base

• Only complete enumeration of the population
  (especially at a person level)
• Uses a similar definition of usual residence to
  the population estimates
• Once a decade the population estimates are
  rebased using the most recent Census
• Many distributions at local level are also based
  on Census data
• Very accurate source

9
Births and DeathsSources

• Birth Registrations
• Provided by Vital Statistics Outputs Branch
  (VSOB) in ONS
• Death registrations
• Also provided by VSOB
• Adjusted from age at death to give age as at
  mid-year (by VSOB)

10
BirthsMethodology

• Births added on to population at age zero.

11
BirthsChallenges

• Assume births abroad births to non EW
  residents
• these births are allocated in proportion to the
  number of females of childbearing age (15-44)
• Births to non-resident mothers is a big issue.
• Allocation of 0 year old migration.
• Unlike deaths (more later) age at birth not an
  issue - not many people are born already aged 30!

12
DeathsMethodology

• Subtract deaths from base
• BUT
• January to June data are not final
• So adjusted for late registrations by
• adding on adjustments for previous year
• assumes adjustments are small
• assumes level of adjustments similar across years
• avoids need for annual revision

13
ChallengesDeaths

• Very accurate source -
• but not perfect
• Late Registrations (some now included)
• Incorrect date of birth recorded
• Deaths abroad vs. Deaths of non residents
• Deaths Elsewhere allocated in proportion to
  other deaths
• Residence issues

14
Internal MigrationDefinitions

• Definition
• A person who changes their place of usual
  residence within the United Kingdom
• Internal migration is measured by looking at GP
  re-registrations
• An Internal migrant is someone who
• notifies their current GP of a change of address
  or
• registers with a new GP in a different area from
  their previous GP

15
Internal MigrationSources 1

• National Health Service Central Register (NHSCR)
• Moves between former HAs
• General practitioners (GPs) patient registers
• Lower geography but only annual download
• Points to note
• Reliant on registration with doctor
• Possibly differing propensity to register
• Time period is end July to end July
• (one month delay in registering allowed)
• Armed forces and (currently) prisoners excluded

16
Internal MigrationSources 2

• Combining the two sources
• Estimate moves between former health authorities
  for those aged one and over, from both sources
• Derive scaling factors to apply to patient
  register data
• Derive estimates for under-1s from estimates for
  one year-olds

17
International MigrationDefinitions

• International MigrationCivilian migration
  between the UK and the rest of the world
• International MigrantA person who changes his or
  her country of usual residence for a period of at
  least a year

18
International Migration Overview 1

• Total International Migration (TIM) estimates
• International Passenger Survey migration data
• Migrant Switcher adjustment
• Visitor Switcher adjustment
• Asylum Seeker adjustment
• Irish flows

19
International Migration - Overview 2
Total International Migration
20
Irish (Eire) Flows

• The International Passenger Survey (IPS) does not
  cover people crossing the Northern Ireland/
  Republic of Ireland border
• Therefore, an estimate of migrants to and from
  the Republic of Ireland needs to be made, using
  another data source
• To avoid double-counting, any of these migrants
  captured by the IPS are removed from the IPS
  component

21
Irish (Eire) FlowsSources

• Data supplied by Migration Statistics Unit
• by country gender
• provided by CSO(Ireland)
• based on responses to the Irish Quarterly
  National Household Survey (April)
• These estimates cover
• In migrants to Eire from EW in the last 12
  months
• Out migrants to EW who were living in Eire 12
  months ago

22
Irish (Eire) Flows Methodology 1

• Split country totals to HA level
• - Using 2001 EW Census Distributions
• Inflows previous address 1 year ago in Eire
• Outflows country of birth Eire
• Split HA totals to LA level
• Inflows previous address 1 year ago in Eire
• (2001 EW Census)
• Outflows previous year resident population

23
Irish (Eire) Flows Methodology 2

• Apply age-sex distribution
• Using data from the EW and Irish Censuses
• National age-sex distribution used
• Inflows 2001 EW Census
• Outflows 2002 Irish Census

24
Asylum Seeker Adjustment Definition

• Inflows are those coming into the UK claiming
  asylum
• who are granted asylum or
• whose case will take 12 months or more to process
• Outflows include people removed people who
  have withdrawn their application and state their
  intention to leave and an estimate of the
  proportion of failed asylum seekers who leave
• The asylum seeker adjustment only includes
  asylum seekers missed by the IPS

25
Asylum SeekerSources

• Migration Statistics Unit (from Home Office
  data)
• AS inflow and outflow data for Principal
  Dependent applicants
• UK by sex quinary age groups
• GORs
• Home Office
• NASS (National Asylum Support Service)
• Provides figures on subsistence and accommodation

26
Asylum SeekerMethodology

• Using the NASS data
• For each LA calculate the number of AS in receipt
  of an allowance (either accommodation and
  subsistence or subsistence only)
• Calculate the distribution within each GOR.
• Also calculate the distribution for subsistence
  only recipients.
• Assume 70 of AS applicants (and their
  dependents) in receipt of an allowance and the
  remaining 30 have a similar distribution to
  those receiving subsistence only

27
Asylum SeekersChallenges

• Definition difference between Population
  Estimates and National Population Projections.
• Use successful Asylum Seekers
• Failed Asylum Seekers Simply not sure where they
  go is proportion correct?
• Distributions
• Use of National age-sex distribution at LA level
• Expanding quinary age groups using equal spread/
  85 cut-off

28
International Passenger Survey (IPS)

• Continuous voluntary sample survey
• Conducted by ONS since 1964
• Tourism, balance of payments and migration
• Principle air, sea Channel Tunnel routes
• Approximately 0.2 of all travellers sampled -
  immigration filter shifts increase sample size
• Based on intentions for Length of Stay

29
IPS - Length of Stay Question
Visitor data
30
International Migration (IPS)Sources

• Data provided by Migration Statistics Unit
• by age sex
• by HAs
• High Weights data
• High Weights
• Where 1 IPS contact is grossed to an excess of
  1000.

31
International Migration Methodology 1

• Calculate GOR Totals
• - Aggregate IPS data (incl. high weights) to GOR
  to produce GOR totals
• Calculate an HA distribution (excl. high weights)
• - Subtract high weights from IPS HA totals to
  produce HA distribution for each GOR
• Apply HA distribution to GOR total to produce
  new HA totals

32
International Migration Methodology 2

• The HA totals are then smoothed over 3 years
• 3yr HA total x GOR total
  (current year)
• 3yr GOR total
• Split by LA within HA
• Inflows using Census distribution usual address
  1 year ago (all immigrants)
• Outflows using previous years population
  distribution
• Apply National age-sex distribution (IPS)

33
International MigrationChallenges

• Overlap with Armed Forces
• Particular areas difficult to estimate.Outflows
  for these areas may result in an underestimation.
• First Onward Moves
• Links between Internal and International
  Migration. May be missing vital data.
• Sample Variation
• Intentions vs. Actuality

34
Visitor and Migrant Switchers Definitions

• Migrant Switchers (Subtracted)
• Persons who state the intention (in the IPS) to
  stay in the destination country for more than a
  year but who actually leave sooner.
• Visitor Switchers (Added)
• Persons who enter (or leave) the UK as visitors,
  but subsequently extend their visit to 12 months
  or longer.
• Includes both EEA and non-EEA people

35
School Boarders Sources

• England - DfES (Department for Education and
  Skills, Pupil Statistics Unit)
• Wales - WAG (Welsh Assembly Government,
  Statistical Directorate)
• Collected on annual returns from schools.
• Includes public and state schools.

36
School BoardersMethodology

• Compare previous year figures with current year
  to derive annual change figure.
• Add annual change to population base.
• Schools allocated to LA based on postcode.
• Information on boarders collected at SYOA, 7 to
  18.

37
School BoardersChallenges

• Data Quality
• Data on 81,500 boarders in 860 schools received
• Final Total 82,200 boarders from 920 schools
• Missing Schools (forms not returned)
• Schools incorrectly record day-pupils as boarders
• Is Definition of boarder suitable?
• In recent times more schools have required local
  GP registration - hence boarders may be captured
  through standard internal migration.

38
PrisonsSources

• Home Office

39
PrisonsMethodology

• Compare previous year figures with current year
  to derive annual change figure.
• Add annual change to population base.
• Prisons allocated to LA based on postcode.
• Information on prisoners collected at SYOA, 14 to
  90

40
Prisons Challenges

• Quality of raw data
• Prisons codes can be incorrect in source
• Combining of Prison medical service into NHS.

41
Home Armed Forces (HAF)Methodology 1

• Mid-year population estimates include all people
  usually resident in England Wales but NOT HM
  Forces (or their dependents) stationed overseas
• Data on HM Forces (incl. Gurkhas) received from
  Defence Analytical Services Agency (DASA)

42
Home Armed ForcesMethodology 2

• Each year
• Remove previous year HAF resident in EW
• Add back in all previous year UKAF from EW
• Add previous year EW service families overseas
• Age on, then
• Remove current year UKAF from EW
• Add in HAF resident in EW
• Remove EW service families overseas

43
Home Armed Forces Methodology 3 UKAF from EW

• Receive UKAF by age (16-59) sex
• Apply Census ratio to estimate those from
  England, and from Wales

44
Home Armed Forces Methodology 4HAF resident in
EW 1

• Receive HAF by sex by base LA
• take average of Oct, Jan, Apr July figures
• Re-distribute Whitehall Warriors (These are
  armed forces personnel based in London doing
  desk jobs)
• Remove a percentage (based on 1998 data) from
  Westminister
• Re-distribute across London using Census data

45
Home Armed Forces Methodology 5HAF resident in
EW 2

• Add in Gurkhas
• Receive Gurkhas by unit (all male)
• Assign to base LA
• Use base-to-residence matrix from the Census
  assign to LA of residence (by sex)
• Base county to resident LA matrix
• For each county we only include LAs with more
  than 30 males or 10 females resident except where
  there is a small base

46
Home Armed Forces Methodology 6HAF resident in
EW 3

• Apply age distribution by sex by LA from the
  Census
• Constrain sub-national age-sex totals to national
  age-sex totals provided by DASA

47
Home Armed ForcesMethodology 7Service Children
Overseas 1

• Receive UKAF married men overseas by force
• Apply scaling factors (based on 1998 data) by
  force to estimate children overseas
• Apply Census ratio to estimate those from
  England, and from Wales
• Assume equal number of male and female children
• Apply age distribution

48
Home Armed ForcesMethodology 8Service Children
Overseas 2

• Age Distribution
• Receive estimate of number of children in
  families overseas receiving child benefit by age
• Apply boarding school factors (based on Census)
  to remove those still at school in EW (ages
  11-14)

49
Home Armed ForcesMethodology 9Service Wives
Overseas

• Receive UKAF married men overseas by force
• Apply proportions (based on 1998 data) by force
  to estimate wives overseas
• Apply Census ratio to estimate those from
  England, and from Wales
• Assume wives have same age distribution as
  husbands and apply age distribution for UKAF
  married men overseas.

50
Home Armed ForcesSources 1

• DASA (Defence Analytical Services Agency)
• UKAF by age (16-55) and sex
• HAF by age and sex and base LA
• Gurkhas by unit
• UKAF married men overseas by age and sex
• Her Majestys Revenue and Customs (HMRC)
• Children in families overseas receiving child
  benefit by age (based on 5 sample)

51
Home Armed ForcesSources 2

• Census Distributions
• UKAF from England, from Wales
• Distribution of Whitehall Warriors
• Percentage of UKAF children (aged 11-14) at
  boarding school
• Base-to-residence matrix
• HAF by age and sex and LA

52
Home Armed ForcesChallenges 1

• New data are now available on base LA age
  distributions how do we use these data?
• Is the adjustment for Whitehall Warriors
  required?
• No residence based data currently available from
  DASA base-to-residence matrix may not reflect
  changes in living patterns over the decade

53
Home Armed ForcesChallenges 2

• Use of base-to-residence matrix
• County to LA or LA to LA
• Use of 1998 scaling factors/ proportions in
  service families abroad
• Assumption that wives have the same age
  distribution as their husbands
• Currently, UKAF from EW and Service families
  overseas components only included at national
  level

54
Foreign Armed ForcesMethodology

• Compare previous year figures with current year
  to derive annual change.
• Data include forces AND dependants.
• Specific adjustment made as most people travelled
  on military flights that are excluded from the
  IPS
• IPS data excludes military personnel on
  business
• Age and sex breakdown is based on Census.
• Location information is based on postcode/
  address.

55
Foreign Armed ForcesSources

• Single point of contact is the United States Air
  Force, they collate from
• United States Air Force 20,600
• United States Navy 1,000
• United States Army 700
• Census for distributions.

56
Foreign Armed ForcesChallenges

• Data quality - sometimes invalid or missing
  postcode
• No gender or age breakdown
• so Census age distribution gets progressively out
  of date
• If new LA, with previously no personnel, then
  what distribution should be applied?
• Dependent on goodwill of USAF - data supply
  issues and timeliness
• Security considerations

57
Improving Migration Population Statistics (IMPS)

• Briony Eckstein Kanak Ghosh

58

• Background to the Project
• Overview of the Projects Objectives
• Overview of the Areas of Research
• More Detail
• Local Authority Case Studies
• International Migration
• Internal Migration
• Administrative Data Sources

59
Background 2001 Mid-Year Estimates
53.5m
53,225,000
Original inter-censal discrepancy of 1.1 million
52,084,000
51.5m
Rolled forward from 1991
Based on 2001 census
60
Background 2001 Mid-Year Estimates
53.5m
- 351,000
- 305,000
Remaining unexplained difference of 209,000
275,000
51.5m
Rolled forward from 1991
Based on 2001 census
Revised 1991 post-census adjustment
International migration revisions
2001 post-census adjustments
61
Background Possible Weaknesses
Did the 2001 census under- or over-estimate the
population?
CENSUS BASE
Did the 2001 census incorrectly apportion the One
Number Census adjustments across local
authorities within census Estimation Areas?
Are there births that do not get registered?
BIRTHS
Are there births registered where the baby should
not be counted in the population?
Are there deaths that are not registered? Are
there deaths which are registered in the wrong LA?
DEATHS
Are there people who move but do not register in
the new area?
Are there adults who never register, so remain
counted where they were in full time education?
INTERNAL
MIGRATION
Are there people who move but the move is
incorrectly counted because they register in the
new area but were not registered in the old area?
Are asylum seekers coming in correctly allocated
to LAs?
Are asylum seekers who leave correctly subtracted
from LAs?
How are asylum seeker centres accounted for?
INTERNATIONAL
Are visitor switchers correctly allocated to LAs?
MIGRATION
Are other international migrants coming into the
UK correctly allocated to LAs?
Are other international migrants leaving the UK
correctly subtracted from LAs?
Are flows to and from Ireland correctly
apportioned to LAs?
DIFFICULT /
Armed Forces
School Boarders
Prisoners
Students
MOBILE GROUPS
Are sex ratios plausible?
NEW BASE
Are age / sex percentages plausible?
62
Overview IMPS Key Objectives

• To improve the quality and reliability of
  migration and population statistics
• To ensure that the gap between the rolled forward
  mid-year estimates and the 2011 census results is
  not as large as was the case in 2001
• To ensure that ONS more fully understands the
  causes of/reasons for any gap that does occur in
  2011

63
Overview IMPS Work Streams

• Population Base
• what population bases do users need?
• Alternative Data Sources
• nationally held data
• locally held data
• International Migration
• Internal Migration

64
More detail IMPS Work Streams

• Local Authority (LA) Case Studies
• International Migration
• Internal Migration
• Administrative Data Sources

65
Local Authority (LA) Case Studies

• Investigate whether there are any locally held
  data sources which could be used to produce or
  quality assure the population statistics
• The selection of case study areas
• identify risk factors
• group LAs into clusters
• principal component analysis to identify clusters
  of interest
• LA nearest to centroid of each cluster selected

66
LA Case Studies Risk Factors

• Gross international migration
• Gross internal migration
• Armed forces
• Non-white population
• Full-time students aged 18-74
• Prisoners
• Men aged 20-39
• Population density

67
LA Case Studies Selected Clusters LAs

• Cluster 1
• high population density, migration rates and
  proportion of young males
• above national average proportions of students
  and non-white population
• Cluster 2
• high proportion of non-white population
• above national average population density,
  migration rates and proportions of students and
  young males
• Cluster 8
• above national average proportions of non-white
  population, students and young males
• Cluster 11
• high proportion of students

68
International Migration Work Streams

• National Statistics Quality Review research
• International Passenger Survey (IPS) - migration
  questions and sampling
• methods used to integrate sources currently used
  to estimate international migration
• methods used for distributing IPS flows
  geographically within the UK
• potential use of existing alternative data
  sources to estimate international migration
• Other international migration research
• short-term international migration

69
International Migration
Distribution of inflows within the UK, 2001
70
International Migration
Distribution of inflows to LAs within GORs
71
Internal Migration

• Research areas
• alternative data sources
• benefits data for children and older persons
• school census data for children
• student data
• the under-recording of young male migration
• is the migration of other groups under-recorded?
• ethnic minorities
• non-English speakers
• the introduction of walk-in centres
• the closure of GPs lists
• the assumptions for armed forces and their
  dependants

72
Administrative Data Sources

• Research the potential use of
• National governmental sources
• schools census
• national insurance
• child benefit
• older persons
• electoral roll
• Local governmental sources
• council tax
• housing needs surveys
• Commercial sources
• utility companies
• commercial demographic databases (eg. ACORN)

73

• Background to the Project
• Overview of the Projects Objectives
• Overview of the Areas of Research
• More Detail
• Local Authority Case Studies
• International Migration
• Internal Migration
• Administrative Data Sources

74
Contact Details
Website http//www.statistics.gov.uk/imps E-mail
imps_at_ons.gov.uk Telephone Briony (01329)
813672Kanak [phone]