Household Projections for England Yolanda Ruiz DCLG 16th July 2012 Overview  What are household projections  Key data inputs  Methodology Overview (used for 2008-based.

Download Report

Transcript Household Projections for England Yolanda Ruiz DCLG 16th July 2012 Overview  What are household projections  Key data inputs  Methodology Overview (used for 2008-based.

Household Projections for England
Yolanda Ruiz
DCLG
16th July 2012
Overview
 What are household projections
 Key data inputs
 Methodology Overview (used for 2008-based update)
 Stage 1
 Stage 2
 Main outputs & dissemination
 Next update
What are household projections
 Household projections show the long-term trend (25
years) in household numbers if previous demographic
trends were to continue.
 They are not forecasts as they do not attempt to predict
the impact that future government policies or other
factors might have on demographic trends.
 Household projections are a key part of the evidence
base on housing requirements.
 The latest set of projections (2008-based) was published
in November 2010.
Key data inputs
Key data inputs
National and sub-national ONS population projections, by sex
and five-year age bands
Marital status projections
Institutional population (constant at 2001 levels for under 75’s
and share constant for over 75s)
Census data
LFS data
Methodology overview
Methodology overview
Methodology used for first time in last update (2008-based).
The methodology adopted has a two stage approach:
 Stage 1 uses population projections and projected
household representative rates to project household
population by age, gender and relationship status
 Stage 2 uses data from the 1991 and 2001 Censuses to
disaggregate into detailed household types
Methodology overview- STAGE 1
Stage 1:
•
National and sub-national population projections are trend based and
take into accounts assumptions on future levels of fertility, mortality and
migration based on observed levels over a 5-year period
•
ONS marital status projections are used to give 3 relationship groups:
- Couples (married and cohabiting, where they live together)
- Previously married: i.e. separated, divorced or widowed
- single (never married)
•
Institutional population is subtracted from total resident population
projections by 5 year age band, sex and marital status, obtaining private
household population
•
The number of household is essentially the private household population
multiplied by the appropriate household representative rate.
Methodology overview- STAGE 1
Projecting Household Representative Rates
•Household representative rate= is the probability of anyone in a particular
demographic group being part of a separate household (value between 0 and
1)
•Household representative rates are projected forward using a combination of
two fitted trends: a simple logistics and a dampened logistics trend, with LFS
data incorporated into projections for the 2002 to 2009 period.
•
Simple logistics trend- trend fitted to ln (Xt / (1-Xt))
•
Dampened logistics trend- S-shaped curve is fitted to ln (Xt / (1-Xt))
•Local authority totals are constrained to regional totals, and these to England
totals.
Methodology overview- STAGE 2
Stage 2:
•It disaggregates Stage 1 household projections at LA level to 17 detailed household
types.
•Uses 1991 and 2001 census data at LA level by household type and age group of the
head of household to generate draft ‘headship rates’ and ‘non-headship rates’.
•Headship rates= the number of people who head a particular household type within a
particular age group and local authority area.
•It combines three different definitions of household representative:
-oldest male then the oldest female if there is no male (used in previous
methodology)
-1991 Census definition – first named person on the Census form
-2001 Census definition – eldest economically active person then the oldest inactive
person
HOUSEHOLD TYPES
One person households
Male
Female
Couple: No dependant children
Couple: 1 dependant child
Couple: 2 dependant children
One family and no others
Couple: 3+ dependant children
Lone parent: 1 dependant child
Lone parent: 2 dependant children
Lone parent: 3+ dependant children
No dependant children
A couple and one or more other adults
1 dependant child
2 dependant children
3+ dependant children
1 dependant child
Lone parent and one or more other adults
2 dependant children
3+ dependant children
Other households
Methodology overview- STAGE 2
Stage 2…continued:
•The draft headship and non-headship rates are extrapolated forward using a
two-point exponential method.
•Projected draft headship rates are constrained to sum to 1 within each age
band and local authority district.
•Draft household projections are calculated by applying projected headship
rates to the household projections by age band, constraining these to number
of households by local authority from Stage 1.
•Constrained draft households go through minimum adults checks and
dependent children checks before becoming final and aggregated to GOR
and England levels.
England total
population
by sex, age,
marital
status
Institutional
population,
England
Control GOR
to ENG
Control
GOR to
ENG
Control LAD
to GOR
Control
LAD to
GOR
Import
Census
years
HRR
Household
population
Prison
population
adjustment
Aggregate
categories
Aggregate
age-bands
Adjust for
household
representativ
e differences
Two point
extrapolation
of headship
rates
Small count
regional
adjustment
Draft
headship,
nonheadship
rates
Constrain to
sum to 1
within ageband and
LAD
Constrain to
stage one
households,
LAD
Draft
household
s by type,
LAD
HRR
dampened
logistic
trend
projection
Stage one
households
, GORs
Weight
HRR
simple
and
dampened
trends
Stage one
households
, England
LFS
HRR
adjustmen
t
Minimum
adult check
Minimum
adult
correction
Min
dependent
children
check
Min
dependent
children
correction
Aggregate
GOR
household
types to
England
Aggregate
LAD
household
types to
GORs
Stage two
S2 START
Import
LAD
Census
household
type
tables
Adjust all
pensioner
household
s
Stage one
households,
Local
Authorities
HRR
simple
logistic
trend
projection
Import
LFS
HRR
Stage one
Import
ONS
population
projection
s
S1 START
Import
ONS
marital
status
projection
s
Main outputs & dissemination
Main Outputs
• Main projections for England, the regions and local authority districts.
• Variant national projections based on variant population projections published by
ONS. These allow to assess the sensitivity of the household projections to
alternative fertility, mortality and migration assumptions.
• Back projections.
• Components of household growth.
Dissemination
• National Statistics Release.
• Live Tables for England, regions and by district, including number of households
and household population by household type and age of household reference
person. Also tables by number of dependent children and older people.
• Detailed unrounded data for further analysis.
• Methodology report & detailed report by contractors
Next update
Next set of projections – 2010-based, taking into account:
•
•
2010-based national & sub-national main & variant population projections
Labour Force Survey data from 2010 onwards.
Use of 2011 Census data:
•
Today 2011 Census population and household estimates become available. In Autumn,
there are also planned to be ‘short-run’ (5-year) population projections based on 2011
Census. We are considering how to best incorporate this information into our 2010based update.
•
The next set of long-term sub-national projections would be consistent with the new
Census, and these would then be taken into account for the 2012-based update of the
DCLG projections in Autumn 2014.
•
With the new Census data we will be able to assess the differences between projected
2011 household numbers and the 2011 Census results, providing a good diagnostic of
the projection method. It may then lead to a review/ update of the projection method.
Any questions?