Sie sind auf Seite 1von 19

London local authority

household composition
estimates from administrative
data
Dr Gill Harper
Barts Institute of Population Health Science, QMUL
Mayhew Harper Associates Ltd
Understanding Households in Administrative Data: Definition and Estimation
13th March 2018
Background
• First use of linked PCT and LA administrative data to
profile population and health needs in LB Brent 15
years ago
• First administrative data population estimation for
LB Brent 2004 as an alternative to 2001 Census
• Since then over 60 projects in over 20 local
authorities
• Comparison for 2011 Census for the 6 Olympic
boroughs
• Household typology since 2008
Why households?
• Self contained units of production and consumption
• Transactional units for different services e.g. utilities,
some welfare benefits
• Type and demographics of household strongly influence
demand for services
• Characteristics have powerful influence on health and
well being and social outcomes
• Economic contribution of households unrecognised but
estimated to add 40% to GDP
• Increasingly required in different policy domains
Available household statistics
• ONS census survey of population and households
• Every 10 years, OA lowest geography
• DCLG household projections
• Every 2 years, LA level, based on past
demographic trends, ONS popn projections and
Census household composition
• Surveys e.g. GHS
• Samples, for national research and policy
• Commercial geo-demographic
• At cost, consumer and lifestyle typologies
User needs
Local authorities:
• Timely and granular socio-economic information
at sub-local authority level
• To inform policy and services and understand
local context
• Evidence for Localism Act (2011) and Health and
Social Care Act (2012)
Limitations of available sources
• System is fragmented and statistics that are for
specific purposes (HBAI, household surveys)
• DCLG projections are synthetic estimates based on
population forecasts, prescribed household types
• Information tends to be inflexible and unavailable
below local authority level and heavily dependent
on out of date census information
Local administrative data alternative
• Based on local administrative
data population counts
• Links together multiple
sources to improve the
coverage of outputs and
applies rules and
assumptions to determine
who is a current resident
• The ‘minimum confirmed
population’
Admin data population count

A: on any other database


B: on GP register
C: assigned a UPRN (living at a
recognised address)

Venn
eleme decisi
nt a b c on comment
0 0 0 0 R not on any data set
1 1 0 0 R on the GP register only
2 0 0 1 R empty property
3 0 1 0 R on other data set only
4 1 0 1 A on GP and address register
5 1 1 0 R on GP register and other data set
6 0 1 1 A on other data set and on address register
on GP register and other data set and address
7 1 1 1 A register
Summary of stages
QA checks
• Child Benefit numbers published by HM Revenue and
Customs for children aged 0-16
• State Pension claimants by males (65+) and females (60+)
• Comparing the vacant UPRN rate with a local authority’s
own figures or Council Tax records
• UPRNs with high occupancy levels, greater than 9 people,
are identified and checked for being multiple-occupancy
• Comparison with other sources from contemporaneous
snapshots e.g. ONS MYEs or GLA figures, if the local
authority is situated for example in the London area
• Number of children aged <16 without an adult at a UPRN is
checked for possible explanations (e.g. parent or guardian is
not on the GP register).
Compare admin data v official
household counts

• Uses occupied UPRNs from population estimation


• Excludes communal establishments based on LLPG
classifications
• 6 Olympic boroughs 2011
Recreate official household typology
Household type
One person households
Description
Male
• Cannot differentiate between
Female married or unmarried households
One family and no others
a b c
Couple : No dependent children • Identified dependent children from
Couple: 1 dependent child
School Pupil Census and Connexions
Couple: 2 dependent children • The DCLG categories of ‘couple’ or
Couple: 3+ dependent children
‘lone parent’ with one or more
Lone parent: 1 dependent child
Lone parent: 2 dependent
other adults’ were combined for
children simplicity into ‘family households
Lone parent: 3+ dependent
children with other adults and dependent
A couple and one or more other
d
No dependent children children’
adults
1 dependent child
• 2011 Census household types
2 dependent children
3+ dependent children
grouped in the same way as DCLG
Lone parent and one or more 1 dependent child categories for comparison purposes
other adults
2 dependent children
• DCLG 2010 same sex couples not
3+ dependent children
Other households
e
See notes
included. Census 2011 and DCLG
2013 same sex couples are a family.
Compare household typologies
• Similar totals
• Admin fewer ‘couple
family’ and ‘other’
households
• Admin more ‘one
person’ and ‘family
with other adults’
households
• Definitional issues
• UPRN v household
• Usually resident v
service using
population
Overview
• Official household counts have fallen into line with
earlier admin counts
• Discrepancies between (and within) ONS, GLA and
admin sources illustrates how figures can get out of
kilter in areas of high in-migration and regeneration
e.g. Newham
• Local administrative data able to replicate DCLG
typology
• Official sources subject to frequent change in
definitions and population base
Flexible household typologies
• DCLG typology is fixed
• LAs need to identify a range of different households
e.g. pensioner and 3-generation households for
policy
• Use age and occupancy for every address from
administrative data population estimation database
as starting point
• Snapshot in time rather than projections
8 nkm household categories
household
case states 0-19 20-64 65+ household size type
1 0 0 0 0 void category Description
2 a 1 0 0 1 H
3 b 0 1 0 1 G A family households with dependent children
4 c 0 0 1 1 D
5 aa 2 0 0 2 H
B single adult households with dependent children
6 ab 1 1 0 2 B C older cohabiting households
7 ac 1 0 1 2 H
8 bb 0 2 0 2 F D older person living alone
9 bc 0 1 1 2 C
10 cc 0 0 2 2 C E three generational households
11 aaa 3 0 0 3 H
12 aab 2 1 0 3 B F cohabiting adult households no children
13 aac 2 0 1 3 H
14 abb 1 2 0 3 A
G single adult households
15 abc 1 1 1 3 E H other households
16 acc 1 0 2 3 H
17 bbb 0 3 0 3 F
18 bbc 0 2 1 3 C
19 bcc 0 1 2 3 C
20 ccc 0 0 3 3 C
21 aaaa 4 0 0 4 H
22 aaab 3 1 0 4 B
23 aaac 3 0 1 4 H Enumerating households is a combinatorial
24 aabb 2 2 0 4 A
25
26
aabc
aacc
2
2
1
0
1
2
4
4
E
H
problem. Our starting point is the population
27
28
abbb
abbc
1
1
3
2
0
1
4
4
A
E which we then split into three different age
29 abcc 1 1 2 4 E
30 accc 1 0 3 4 H groups. According to the membership of each
31 bbbb 0 4 0 4 F
32
33
bbbc
bbcc
0
0
3
2
1
2
4
4
C
C
household each is assigned to one of the eight
34
35
bccc
cccc
0
0
1
0
3
4
4
4
C
C
basic categories. However, this can be varied
by using e.g. more age categories or gender
Integrate data with UPRN
Number of
children in
household number of % social % on
age 0-19 households housing benefits
void 5975 0 0
0 73,684 45.7 33.3
1 13,732 60.2 48.5
2 9,630 63.9 53.5
3 4,420 71.1 61.7
4 1,803 71.2 66.5
5 702 63.8 71.2
6 355 53.8 73.2
7 188 38.8 77.1
8 84 32.1 79.8
9 12 33.3 83.3
10 1 100.0 100.0
>10 41 9.0 24.0
Grand total 110,627 48.3 37.5

• Link households to other attributes


• Tenure
• Means tested benefit claimant
• Levels of access to services and amenities
• To address specific issues of interest
• Workless households
• Households with complex needs
• Users of council services
Conclusion
• Admin data provide a comparable count and typology at the
local level with more timeliness, flexibility, granularity and
linkage
• ONS ADC similar concepts: linking multiple sources with
different coverage, applying rules, associative matching
techniques, ‘superseding’ assumption rules, activity data to
identify most recent address, concept of ‘occupied
address/UPRN’
• Differences: uses Census 2011 as gold standard, accessible
data sets, PCS to develop confidence intervals, anonymised
linkage, ‘usual resident’ population, nationally comparable,
national statistics standard, more stakeholders
• Easier to make radical changes at the local level
Harper, G. and Mayhew, L., 2016. Using Administrative Data to
Count and Classify Households with Local
Applications. Applied Spatial Analysis and Policy, 9(4),
pp.433-462.

g.harper@qmul.ac.uk

Das könnte Ihnen auch gefallen