28th june, 2017 dami aguda (heriot-watt university) · 2017-06-22 · 28th june, 2017 dami aguda...
TRANSCRIPT
28TH JUNE, 2017 DAMI AGUDA (HERIOT-WATT UNIVERSITY)
Homeownership Transition Influences among British Young Adults: Do Neighbourhood
Effects Matter?
Research aim is to explore socio-psychological influences of housing tenure choice among British young adults. Presentation outline: * Introduction * Theoretical background * Results * Conclusions: Implications of the research and future work
Introduction
Homeownership: The largest share of the three major housing tenure in Britain
Source: Office for National Statistics (2016)
0
10
20
30
40
50
60
70
80
% Housing stock by tenure in Great Britain (1951-2015)
Owner occupied Private renting Social renting
Introduction
First-time buyers: Predominantly young adults Source: Nationwide (2016)
Introduction
0.0
2.0
4.0
6.0
8.0
10.0
12.0
198
3
198
4
198
5
198
6
198
7
198
8
198
9
199
0
199
1
199
2
199
3
199
4
199
5
199
6
199
7
199
8
199
9
200
0
200
1
200
2
200
3
200
4
200
5
200
6
200
7
200
8
200
9
2010
2011
2012
2013
2014
2015
2016
Rat
io o
f h
ou
se p
rice
to
ear
nin
gs
Regional house price to earnings ratio for first time buyers
London Wales Scotland N Ireland UK England
What are the suggestions?: Q1. Are homeownership decisions influenced by intergenerational assistance through inheritance/financial expectations? Q2. Are homeownership decisions influenced by path dependency through parental/family motivation or by societal/local norms? Data sources and methods:
*Review of literature
*Summary statistics and cross-tabulation of data from:
-HM Revenue & Customs
-English House Conditions Survey
-Wealth and Assets Survey
Q1 *Principal components analysis
*Fixed-effects regression
*Multi-level binary logistic regression
Using data from:
-Index of multiple deprivation
-British census data
-British Household Panel Survey
-Office for National Statistics
Q2
What are the established drivers of housing tenure decisions in Britain?: Build-up of the socio-psychological context
Econometric/theoretical context
Economic
-Household income
-House prices
-Employment status
-Cost of owning to renting
Demographic
-Age
-Sex (male/female)
-Household formation
-Race
-Geographic mobility/location
Other theoretical
context
Socio-psychological
-Beliefs/expectations
-Norms
-Values
-Socialisation
-Status/wellbeing
-Motivations
-Family background
Attitude towards homeownership
Norm/standards set by (network of)
trusted individuals
Perceived behavioural
control
Intention (not) to own
Actual decision (not)
to own
Adapted from Ajzen, I. 1991.The theory of planned behaviour
Theoretical background
Non-spouse related housing estates passing on death, 1999-2014
Year Housing Estates Value as % of all
property assets
Value as % of total
assets
Number
(thousands)
Value
(billion £)
1999-00 107.94 9.85 64.85% 25.53%
2001-02 107.40 12.51 62.78% 27.40%
2002-03 115.95 15.88 66.81% 32.69%
2003-04 121.50 18.46 68.92% 35.59%
2005-06 123.49 20.07 65.82% 35.21%
2006-07 121.46 20.81 66.27% 35.19%
2007-08 115.40 20.62 64.61% 34.45%
2008-09 114.05 18.75 64.35% 32.25%
2009-10 118.13 18.68 63.77% 32.81%
2010-11 124.83 19.21 65.27% 34.28%
2011-12 132.20 19.63 66.70% 34.60%
2012-13 138.45 20.55 66.55% 34.52%
2013-14 138.78 22.22 66.30% 35.00%
Source: Table 12.5 of HM Revenue & Customs (2016)
Results
Hamnett, C. et al. (1991)’s evidence from Inland Revenue data: Housing inheritance would become a major source of homeownership over the next four decades from the early nineties. But previous figure suggests that there is no substantial increase in the size of British housing inheritance Other updates: Rowlingson, K. and Mckay, S. (2004) and Hamnett, C. (1997)….both agreed that that the initial prediction is wrong due to growth in private residential care homes for older people; and -Young adults have lower chances of inheriting estates What then are the age groups of beneficiaries of housing inheritance?
Results cont.
Based on the responses from the heads of households :
Source: English House Conditions Survey (EHCS)
2 1 2 0 0 4 6 2 0
5 4 1 2
35 38
30 24 23
10 6
6 7
31 24
21 19 21
10 11
12 11
35 29 13 18 17
11 7 12
15
25 24 17 15 19
7 7 8 3 7 9 10 9 9
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2003 2004 2005 2007 2009 2010 2011 2012 2013
(N = 694)
Home owners (within 5 years length of ownership) sourced from inherited or gifted house or money by age groups in percentages of total population per year (actual
numbers as labels)
16-24 25-34 35-44 45-54 55-64 65+
- Whether home was acquired as inheritance
- Whether home was acquired as/or using gift
Results cont.
Q2 – Work in progress: Firstly, by predicting local area home ownership rates using their corresponding socio-economic and deprivation features with data specific to young adults (aged 16-34) Three-staged procedure: 1. OLS regression ……………….. zi = αi + ϰiβ + ui
*Not presented for brevity
2. Principal Components Analysis (PCA) *For reducing : dimensionality multicollinearity
3. Fixed-effects regression … zit = ϰit1β1 + ϰit2β2 + … + ϰitkβk + uit where: zit represents homeownership rate of a neighbourhood ‘i’ in the year ‘t’ ; ϰit represents principal components (i.e. social characteristics) of the neighbourhood ‘i’ in the year ‘t’ ; αi represents any neighbourhood-specific omitted variable in the estimation; β represents the coefficient estimates; and uit represents the error term of the estimation.
Principal components grouping for census and deprivation data
2001
Pc1 Degree and higher, A-levels, lower O-levels, higher professional, lower
professional, lower technical, semi routine, routine, no dependent child(ren)
Pc2 Higher O-levels, intermediate, small employees, unemployed, couple, Q-IMD5,
Q-Income5, Q-Employment5
Pc3 Q-IMD2, Q-Income2, Q-Employment2
Pc4 Q-IMD3, Q-Income3, Q-Employment3
Pc5 Q-IMD4, Q-Income4, Q-Employment4
2011
Pc1 Degree and higher, lower O-levels,
higher professional, lower
professional, semi routine, routine,
no dependent child(ren)
Pc2 A-levels, higher O-levels,
intermediate, small employees,
lower technical, unemployed,
couple, Q-IMD5, Q-Income5, Q-
Employment5
Pc3 Q-IMD2, Q-Income2, Q-
Employment2
Pc4 Q-IMD3, Q-Income3, Q-
Employment3
Pc5 Q-IMD4, Q-Income4, Q-
Employment4
0
2
4
6
Eig
en
va
lues
0 5 10 15 20 25 Number
Fixed-effects principal components regression of neighbourhood homeownership rate on their socio-economic attributes
*** denotes significance at 1%; ** at 5% and * at 10%
Principal components Coefficient
Standard
error t-statistic
Pc1 3.385 0.099 34.32***
Pc2 1.521 0.121 12.54***
Pc3 -0.348 0.135 -2.59**
Pc4 1.886 0.121 15.53***
Pc5 0.925 0.111 8.35***
Year 2011 -12.515 0.124 -100.93***
Constant 71.168 0.073 973.54***
N 80867
prob>F 0.00
F-statistic 5306.68
R-squared 0.4679***
Results cont.
Implications and future work * YES – Neighbourhood effects really matter. * Local area homeownership rates can be predicted from their corresponding socio-economic features in Britain. * The reduced components can form neighbourhood-level predictors in an impending two-level housing tenure choice model. Future work: * The results obtained are only a demonstration of what to expect when these neighbourhood-level components are combined with individual-level data to pull out neighbourhood effects. * But firstly, an update to the duration analysis of time to homeownership findings in Britain, running up to 23 waves (1991-2015) at the individual level only. * Duration of socialisation in parental homeownership is an additional variable to be included in our two-level housing tenure choice model at the individual level.
Conclusions
Ideas on neighbourhood effects: Following data from related literatures on neighbourhood features as drivers of later outcomes. e.g. Harkness, J. and Newman, S. J. (2003) Effects of homeownership on children: The role of neighborhood characteristics and family income. Economic Policy Review, 9 Bramley, G. and Kofi Karley, N. (2007) Homeownership, poverty and educational achievement: School effects as neighbourhood effects. Housing Studies, 22, 693-721
Additional information
Some deprivation variables and concept: LSOA highest academic qualification (default is ‘no qualification’) LSOA Socio-economic class (default is ‘long-term unemployed’) LSOA unemployment rate LSOA household living arrangement (‘single’, ‘has dependent child(ren)’ as defaults) LSOA IMD (in quantiles (Q), ‘Q1’ is most deprived and default) Where: LSOA: Lower super output area or data zones of 1000-3000 population IMD: Index of multiple deprivation Quantiles: An innovative way to combine the deprivation samples to form a British sample. Quantiles is an equal division of a set of values of a variate frequency distribution, each containing the same fraction of the total population.
Additional information