The political consequences of housing (un)affordability

Abstract

The enormous growth in house prices in Europe since the 1990s has led to increasing concerns about the affordability of housing for ordinary citizens. This article explores the relationship between housing affordability – house prices relative to incomes – and the demand for redistributive and housing policy, using data drawn from European and British social surveys and an analysis of British elections. It shows that, as unaffordability rises, citizens appear in aggregate to become less supportive of redistribution, interventionist housing policy and left-wing parties. However, this aggregate rise, driven by the predominance of homeowners in most European countries, masks a growing polarization in preferences between renters and owners in less affordable regions.

Full text

Article
Journal of European Social Policy
2021, Vol. 31(5) 597–613
© The Author(s) 2021
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/09589287211056171
journals.sagepub.com/home/esp
The political consequences of
housing (un)affordability
Ben Ansell
Nuffield College and Department of Politics and International Relations, University of Oxford, Oxford, UK
Asli Cansunar
Department of Political Science, University of Washington, Seattle
Abstract
The enormous growth in house prices in Europe since the 1990s has led to increasing concerns about the
affordability of housing for ordinary citizens. This article explores the relationship between housing af-
fordability –house prices relative to incomes –and the demand for redistributive and housing policy, using
data drawn from European and British social surveys and an analysis of British elections. It shows that, as
unaffordability rises, citizens appear in aggregate to become less supportive of redistribution, interventionist
housing policy and left-wing parties. However, this aggregate rise, driven by the predominance of home-
owners in most European countries, masks a growing polarization in preferences between renters and
owners in less affordable regions.
Keywords
housing, inequality, affordability, political economy, redistribution, preference
Introduction
There has been a recent surge in scholarly interest in
the political effects of inequality in the advanced
industrial world. However, most existing work has
focused on inequalities in labour market incomes,
neglecting the growing importance of inequality in
wealth and, in particular, housing. As Fuller et al.
(2020) show, changes in house prices are largely
responsible for short- and long-run shifts in wealth
inequality in the industrialized world. In most
wealthy countries there has been a surge in nominal
house prices since 1990, barely broken by the crash
of 2008/9 (Knoll et al., 2017), even though incomes
have stagnated. This housing boom –and the ensuing
growth of wealth inequality –has created winners
and losers. Increases in housing prices, which
translate into increased rents, mortgage payments
and down-payments, often cause severe burden on
many households’budgets (Albouy et al., 2016;
Dewilde, 2018;Dustmann et al., 2018). The flip-side
of growing unaffordability is the emergence of a
large group of ‘winners’: people with modest in-
comes but rapidly appreciating assets, whose welfare
Corresponding author:
Ben Ansell, Nuffield College and Department of Politics and
International Relations, University of Oxford, Nuffield College,
New Road, Oxford, OX1 2JD, UK.
Email: ben.ansell@politics.ox.ac.uk

is increasingly attached to their homes not their jobs
(Ansell, 2014).
One might expect serious political consequences
to housing unaffordability. However, the precise
direction of these consequences is not entirely ob-
vious. On the one hand, people unable to enter the
housing market might seek greater government
support through housing policy or through redis-
tributive transfers. Some scholars viewed the relative
success of Jeremy Corbyn’s Labour Party in 2017 as
reflecting this squeezed ‘Generation Rent’(Ansell
and Adler, 2019). On the other hand, there are clear
beneficiaries of unaffordable housing –those who
own it. Homeowners may seek to protect their
windfalls from direct taxation or the indirect threat of
new construction. Existing work, however, looks
only at house prices in absolute terms and looks only
at redistribution not housing policy preferences. As
yet, we do not have empirical clarity on whether the
surge in housing costs relative to incomes raises or
lowers the demands for such policies. Nor do we
know whether homeowners and renters are reacting
differently to this new environment.
In this article, we argue that housing unafford-
ability plays a key role in shaping both overall support
for redistributive fiscal and housing policy and the
degree of polarization between renters and home-
owners in their preferences. Following Ansell (2014),
we claim that higher housing prices relative to income
produce a net windfall for homeowners that they are
anxious to protect from both taxation and the con-
struction of new housing that might push prices down.
Hence, rising unaffordability leads to an aggregate shift
to less redistributive policy attitudes, more right-wing
voting and greater antipathy to government interven-
tion in housing markets, largely because in almost all
European regions, homeowners form a majority and
want to protect the value of their housing.
This overall ‘right-ward’shift masks underlying
polarization, however. Renters, since they do not
own housing assets, do not benefit directly from rises
in the price of these assets. Less affordable housing
may make them relatively more supportive of the
government intervening in the housing market. Thus,
as unaffordability increases we expect to see the
preferences of homeowners and renters to diverge
substantially. Unaffordability produces polarization.
We test these propositions using a wide variety of
data sources. We begin by examining housing un-
affordability across Europe, drawing data on af-
fordability from the EU-SILC household panel and
matching it at the regional level to policy preferences
in the International Social Survey Program. This
regional data allows us to move beyond existing
work, which largely examines house prices at the
national level or within single countries. We show
that regions with less affordable housing also tend to
have lower support for redistribution and interven-
tionist housing policy and that this effect is mainly
concentrated among homeowners. We then turn to
the United Kingdom to examine both housing policy
preferences and voting outcomes. We start by
looking at how regional affordability conditions the
housing policy preferences of owners and renters,
using the British Social Attitudes Survey. Finally, we
develop original measures of housing affordability at
the parliamentary constituency level and examine
how changes in affordability connect to changes in
support for the Conservative Party. These three
separate analyses permit us to move from how
housing affordability affects individual preferences
to its impact on individual behaviour, as well as
considering the degree to which housing afford-
ability polarizes not only attitudes towards housing
policy itself but also towards the broader role of
government in narrowing market differences.
Argument
Although the boom in property prices relative to
incomes since the 1990s has been one of the core
economic trends of the past two decades, social
scientists have only recently turned to examine
property prices as determinants of political prefer-
ences and behaviour. Early work, beginning with
Kemeny (1981) drew out a potential connection
between housing markets and support for the welfare
state. Kemeny argued that the rise in private
homeownership over the 20th century –and accel-
erating since 1945 –was creating a class of citizens
who would depend less on the state for support,
having a valuable asset to fall back on, and who
would be more tax-sensitive given the costs of af-
fording down-payments and mortgage payments
598 Journal of European Social Policy 31(5)

associated with the purchase of a property. Castles
(1998) and Conley and Gifford (2006) both demon-
strated a cross-national negative relationship between
homeownership rates and the size of government.
The political and social effects of the housing
market are not about ownership alone. Property
prices also matter –both absolutely (as returns to
residential investment) and relative to incomes (in
terms of affordability). Indeed, the major shift of the
past few decades has not been in ownership rates, but
the relative price of housing. Furthermore, Fuller et al.
(2020) show that the effects of housing on wealth
inequality travel through price differences not owner-
ship rates per se. So how might rising housing prices,
relative to income, matter for policy preferences?
Existing work looks at absolute prices rather than
prices relative to income, thereby largely neglecting
housing affordability. Scheve and Slaughter (2001),
in important early work, argue that homeowners in
areas exposed to international trade tend to be less
supportive of free trade as it is likely to push down
the value of their house. Ansell (2014), by contrast,
examines the determinants of domestic social policy
preferences. He argues that rising house prices make
owners less supportive of the welfare state and re-
distribution since increased prices act as a ‘windfall’,
increasing homeowners’permanent income. That, in
turn, makes them more tax-averse and less in need of
social insurance to cover bad times. Hence, owning a
house acts as a form of ‘private insurance’. In this
line of scholarship, researchers look at why the value
of assets such as housing might matter for policy
preferences controlling for labour market incomes.
However, rather than looking solely at the impact of
housing net of income, we also need to consider
housing relative to income.
Relative levels –housing affordability –matter in
a number of ways. We can think of this both in terms
of housing costs relative to income and the value of
housing assets relative to income. Housing is distinct
from most other categories of consumer spending
since it has both consumption and investment value.
On the consumption side, annual housing costs re-
flect how much people are willing to pay for shelter
in various locations and qualities of residence. These
costs might be incurred through rents or through
purchasing a house, typically amortized through a
mortgage. When the costs of shelter in a particular
location or residence rise relative to the income used
to pay these costs, this clearly places pressure on
household budgets. For a renter, whose rent pay-
ments are fully transferred to their landlord, this has
no obvious upside. However, home-owners who pay
their housing costs through a mortgage are not solely
consuming ‘shelter services’they are also investing
in the very asset in which they live. Eventually, once
the mortgage is paid off, they own the asset outright.
And the effects of rising house prices only matter
when they initially take out a mortgage. From their
perspective, higher house prices may produce ini-
tially higher down-payments and monthly mortgage
payments that in part accrue to the mortgage lender,
but they also increase the value of the asset they own.
Relative to income, the former is a burden but the
latter a boon. And over time the latter massively
outweighs the former.
Finally, for those who own their house outright,
they incur no housing costs (presuming imputed
rents are not taxed) and their investment rises in
value. If their incomes stagnate but house prices rise
they are substantially better off. The housing boom
has created legions of dollar millionaires across the
industrialized world. People with relatively modest
incomes have found themselves owning properties
that are several multiples of their annual income in
value. For these people, the ‘affordability crisis’does
not feel like a crisis –it feels like a boon.
So, housing becoming more unaffordable relative
to incomes has countervailing effects on people’s
material circumstances. What does this mean for
their policy and political preferences? In this article
we focus on two sets of attitudes. First, we look at
attitudes to redistribution, which, following Svallfors
(1997),Cusack et al. (2006), and Brady and Bostic
(2015), we use as indicative of a general left–right
economic dimension. Second, we look at attitudes to
government intervention in the housing market in
order to make housing more affordable (by subsi-
dizing either housing construction or providing re-
sources to citizens to help them afford housing).
For homeowners, the story is a simple materialist
one.
1
When house prices rise relative to income, they
become better off since their housing costs remain
flat relative to income while the investment value of
Ansell and Cansunar 599

housing rises. Per Ansell (2014), homeowners will
respondbydesiringlesssocialinsuranceandbybe-
coming more tax-averse (see also Ansell and Adler,
2019;Ansell et al., 2018;Stegmueller, 2013). Hence,
we expect homeowners to become less supportive of
redistributive policies or the parties that support them
when housing unaffordability increases.
Similar forces are at work when we consider the
effect of rising housing unaffordability on home-
owners’attitudes towards government housing
policy. If the value of their asset is rising relative to
incomes, homeowners benefit. Government inter-
vention in housing markets, by contrast, may
threaten the value of this investment. In particular
constructing new houses, by raising housing supply,
pushes down the value of existing houses. Gov-
ernment spending on access to housing for non-
homeowners or renters has a less direct effect on
house prices (and could even boost them) but the
problem with these policies from the perspective of
homeowners is that they are very unlikely to be ben-
eficiaries of such schemes but will have to pay the taxes
needed to fund them. Overall, we anticipate home-
owners will thus oppose government intervention in
housing markets aimed at improving affordability.
For renters, the effects of growing unaffordability
are slightly more complicated. On the whole we
would expect rising house prices relative to income
would drive renters to more redistributive prefer-
ences. Since rents typically rise with house prices, it
may be ever harder for them to pay the rent, pre-
suming that incomes remain flat. It will also make it
more difficult to ever get on the property ladder. For
many renters, these struggles may increase their
support for redistribution and the left-wing parties
that support it –particularly if they are struggling to
make payments. There is a nuance here, though.
Redistribution typically means clawing back larger
shares in taxation from individuals with higher la-
bour market incomes. For higher-income renters,
higher redistribution may be a ‘double whammy’–
losing more of their labour market income and facing
higher relative house prices/rents. For authors in the
tradition of Kemeny (1981) this effect of taxation
making it harder to get on the property ladder is an
important component of the homeownership/welfare
state trade-off. For renters then, the effects of
unaffordability on redistributive preferences may be
mixed. Compared to homeowners, however, they are
clearly less likely to view redistribution more neg-
atively as relative house prices rise.
With regard to housing policy preferences, renters
should have more clear-cut views. Government in-
tervention in housing markets to increase afford-
ability, either by building more houses or subsidizing
their purchase, is largely beneficial for renters –it
reduces the net cost of getting on the housing ladder and
is likely to reduce rents for those who remain tenants.
As local unaffordability rises, the benefits of such
policies should be particularly salient for renters.
2
Putting these two mechanisms together, we expect
that in the medium-run overall rising unaffordability
of housing should lead to less support for redis-
tributive policies and less support for left-wing
parties. Homeownership in almost every European
country hovers over 50% of adults and is typically
higher among voters. Combined with the fact that
homeowners have straightforward anti-redistributive
preferences in the case of rising house prices,
whereas renters may have ambivalent attitudes, we
anticipate that unaffordability will both drive ag-
gregate preferences away from redistributive or in-
terventionist policies and parties and widen the gap
in preferences between homeowners and renters.
These expectations are summarized in Hypotheses
One and Two.
Hypothesis. One: More unaffordable housing should
on average lead to lower support for redistribution,
interventionist housing policies and the (left-wing)
parties that promote them.
Hypothesis. Two: More unaffordable housing should
lead to greater polarization between homeowners and
renters in their preferences about redistribution and
interventionist housing policy.
Housing affordability and attitudes: evidence
from Europe
In this section, we examine the connection between
housing affordability at the regional level and support
for redistribution and housing market intervention
600 Journal of European Social Policy 31(5)

across Europe. To do so we integrate two well-known
existing datasets in an original way. We use the EU-
Survey on Income and Living Conditions (EU-SILC)
to generate our measures of housing affordability by
country and region.
3
We then match this at the regional
level to the International Social Survey Programme for
2006 and 2009. Our interest is in examining the macro-
level effects of living in countries and regions with
different levels of housing affordability on redistribu-
tive and housing policy preferences, as moderated by
homeownership. In other words, how do people in
general –and homeowners in particular –respond to
their housing affordability environment when deter-
mining their attitudes towards redistribution?
Hypotheses One and Two look at how housing
affordability and homeownership structure attitudes
about redistribution and housing policy. To capture
attitudes about redistribution we use the 2009 ISSP,
which asks people whether ‘it is the responsibility of
governments to reduce differences in income between
people with high incomes and those with low in-
comes’. This general redistribution question has been
widely used by political economy scholars seeking to
capture attitudes towards government spending and
the role of the state (Cansunar, 2021;Rehm, 2011),
including existing work on housing and redistributive
attitudes (Ansell, 2014;Ansell et al., 2018). We thus
use this question to provide a general indicator as to
whether housing affordability concerns provoke
general shifts on the standard left–right economic
dimension. The ISSP redistribution question, while
commonly used, does have limitations –countries
start from very different existing baselines of taxation
and redistribution. In order to check our analyses of
the relationship between affordability and economic
left–right measures, in Appendix Tabl e A1 we also
examine a number of related dependent variables –
party choice, taxing the rich and views on inequality.
We also examine the 2006 ISSP, which asks a
question about people’s attitudes towards whether
the ‘government has a responsibility to provide de-
cent housing’. This allows us to directly examine
whether rising housing unaffordability activates
demand for increased or decreased public spending
on housing. Unfortunately, the 2006 ISSP does not
include a measure of individual housing tenure.
However, we use the predicted relationship between
individual demographics and homeownership in
2009 to generate a ‘homeownership propensity’for
2006 in order to separate out individuals who are
more or less likely to be homeowners given what we
know about the determinants of housing tenure.
Our core independent variables are regional
housing affordability and individual home owner-
ship. We begin by setting out the development of our
housing affordability variables. We analyse eight
countries in 2006 (Denmark, Finland, France, Ger-
many, Norway, Spain, Switzerland and the UK) with
58 regions and 11 countries (Austria, Belgium, Den-
mark, Finland, France, Germany, Norway, Spain,
Sweden, Switzerland and the UK) with 70 regions in
2009 using the cross-sectional EU-SILC household
survey. We group households by country and NUTS2
region and calculate statistics on housing affordability
for these region–country–year groups.
4
For each var-
iableweusetheEU-SILC’s survey weights to calculate
weighted averages for each group.
5
We use a variety of different measures of housing
unaffordability at the regional level. We begin with
the average monthly cost of housing in euros at the
regional level for all households (Average Cost). We
split our measure of average costs into those paid on
average by homeowners in that region (Owner’s
Cost) and those paid by renters in that region
(Renter’s Cost). We also look at the regional coef-
ficient of variation in housing costs –the standard
deviation divided by the mean (COV Cost). As
measures of relative cost we look at average housing
cost as a proportion of average gross income (Cost/
Gross Inc) and as a proportion of net income (Cost/
Net Inc.), all measured at the regional level. Figure
A2 provides a graphical presentation of housing
costs in Europe between 2006 and 2017.
6
At the individual level, we include controls for
(logged) household income relative to the mean, sex,
age, urban–rural location, whether the respondent is
employed or retired, education level, union mem-
bership and religiosity.
7
Our main individual variable
of interest is homeownership status. The ISSP 2009
provides this information directly.
8
However, the
ISSP 2006 lacks this variable. In order to proxy for
homeownership when we analyse the ISSP 2006, we
generate a ‘housing propensity’variable. We pro-
duce this by running a regression model predicting
Ansell and Cansunar 601

homeownership in the ISSP 2009 using the same
individual control variables noted above as predictors.
Appendix Figure A1 shows that demographic indi-
cators in the 2009 ISSP are good predictors of home-
ownership in that survey. We then generate predicted
propensities of owning a house in the ISSP 2006 based
on an individual’s characteristics. While this is an
estimate, the estimated average level of homeown-
ership produced in the ISSP 2006 sample is within 2
percentage points of that in the ISSP 2009.
9
Tab le 1
begins with our analysis of how housing affordability
and homeownership affect redistributive preferences.
Throughout our analyses of the ISSP we use multi-
level linear models, with observations at the individual
level, embedded within countries. We have 11 countries
and 70 regions under analysis –with our affordability
variables varying at the regional level in most coun-
tries.
10
We run the models either with random effects at
the country level –the first and third panels –or fixed
effects at the country level –the second and fourth
panels. It is worth noting that those models with country-
fixed effects include only variation in the housing
unaffordability variables across regions in the same
country, whereas the random effects models include
variation at both the country and regional level. The
tables only show the relevant independent variables:
we omit presentation of other individual-level controls.
The top two panels show the direct effects of
homeownership and housing affordability on redis-
tributive preferences and with country random ef-
fects or fixed effects. We see strong evidence of two
things –first that homeowners tend to be less sup-
portive of redistribution in general and second that
regional levels of higher housing costs tend to reduce
support for redistribution. Focussing on Model 5 –
regional housing costs as a proportion of net income –
we see moving from the 10th to 90th percentile on this
variable lowers support for redistribution by 0.66 –over
Table 1. Redistribution preferences 2009 (individual controls not shown).
Random effects
(1) (2) (3) (4) (5)
Average cost Owner’s cost Renter’s cost Cost/gross Inc Cost/net Inc
Homeowner 0.15
∗∗∗
(0.03) 0.14
∗∗∗
(0.03) 0.14
∗∗∗
(0.03) 0.10
∗∗∗
(0.03) 0.11
∗∗∗
(0.03)
Housing variable 1.16
∗∗∗
(0.05) 0.87
∗∗∗
(0.04) 1.21
∗∗∗
(0.08) 4.78
∗∗∗
(0.28) 3.47
∗∗∗
(0.18)
Fixed effects (1) (2) (3) (4) (5)
Average cost Owner’s cost Renter’s cost Cost/gross inc Cost/net inc
Homeowner 0.13
∗∗∗
(0.03) 0.13
∗∗∗
(0.03) 0.13
∗∗∗
(0.03) 0.13
∗∗∗
(0.03) 0.13
∗∗∗
(0.03)
Housing variable 0.89
∗∗∗
0.80
∗∗∗
0.50
∗∗∗
1.53 2.50
∗∗
Random effects (0.20) (1) (0.24) (2) (0.17) (3) (1.52) (4) (1.08) (5)
Average cost Owner’s cost Renter’s cost Cost/gross inc Cost/net inc
Homeowner 0.03 0.01 0.16 0.29
∗∗∗
0.19
∗∗
(0.08) (0.07) (0.11) (0.10) (0.09)
Housing variable 0.96
∗∗∗
(0.10) 0.72
∗∗∗
(0.08) 0.93
∗∗∗
(0.13) 2.92
∗∗∗
(0.55) 2.37
∗∗∗
(0.37)
Homeowner X variable 0.28
∗∗
(0.12) 0.20
∗∗
(0.09) 0.46
∗∗∗
(0.16) 2.50
∗∗∗
(0.64) 1.43
∗∗∗
(0.42)
Fixed effects (1) (2) (3) (4) (5)
Average cost Owner’s cost Renter’s cost Cost/gross inc Cost/net inc
Homeowner 0.08 0.06 0.11 0.00 0.01
(0.08) (0.07) (0.11) (0.10) (0.09)
Housing variable 0.83
∗∗∗
0.72
∗∗∗
0.48
∗∗
0.83 1.92
∗
(0.22) (0.25) (0.21) (1.61) (1.14)
Homeowner X variable 0.09 0.11 0.03 0.87 0.70
∗
(0.13) (0.10) (0.17) (0.64) (0.42)
N9647 9647 9647 9647 9647
Standard errors in parentheses.
*p<0.10, ∗∗p<0.05, ∗∗∗p<0.01.
602 Journal of European Social Policy 31(5)

half a standard deviation. Except in Model 4 we
also find the same pattern for within-country re-
gional differences in affordability –that is, in more
unaffordable regions there is less support for re-
distribution, even netting out the general national
level of unaffordability.
Moving to the bottom two panels we see that the
relationship between higher housing costs (absolute
or relative to income) and declining support for re-
distribution is particularly strong among home-
owners, especially in the random effects models.
11
The left figure of Figure 1 demonstrates this pattern
with Model 5 from the third panel –we see that
homeowners appear to respond to unaffordability
more negatively in terms of redistribution prefer-
ences vis-`
a-vis renters. Note though that there does
appear to be a general negative relationship between
housing costs and redistribution preferences for both
groups and this is robust to both random and fixed
country effects.
As we noted above, redistributive preferences may
not fully account for attitudes along the left–right
economic spectrum. In Table A1 in the Appendix
we examine three different dependent variables that
measure economic left–right preferences –first, vote
choice (a five-point scale increasing to the right,
drawn from the ISSP’s party choice indicator); sec-
ond, support for people with higher incomes paying a
larger share in taxes (a five-point scale increasing in
support); and third, agreeing with the statement that
‘differences in income in [the respondent’scountry]
are too large’(a five-point scale increasing in
agreement). In all cases we examine the fifth model
from the bottom two panels of Table 1–that is the
interactive effect of house prices as a ratio of net
income and homeownership. We see that homeowners
in areas with higher unaffordability are more likely to
support right-wing parties, less likely to think the rich
should pay more in taxes and less likely to think income
differences are too large. Accordingly, we find ample
Figure 1. Regional housing affordability and predicted support for redistribution (left) and housing intervention (right).
Ansell and Cansunar 603

support that the choice of the redistribution question in
Tab le 1 is consistent with wider left–right attitudes.
Table 2 uses the ISSP 2006 to examine the re-
lationship between housing affordability, homeowner-
ship and attitudes towards government responsibility for
housing. Here we have a reduced number of countries –
eight –and 58 regions –again housing affordability
varies in most countries at the regional level.
12
As before,
the top two panels include measures of regional housing
affordability and homeownership (propensity) and the
bottom two panels include the interaction between these
variables. Again, in both cases we look at country
random effects and country-fixed effects models.
The patterns are strikingly similar to what we saw
for redistribution, providing strong evidence that
housing policy preferences and redistribution pref-
erences are driven by similar factors. People pre-
dicted to have a higher propensity to be homeowners
are less supportive of government responsibility for
housing across all nine models. Individuals living in
regions with higher housing costs, whether absolute
or relative to income, are also less supportive of gov-
ernment intervention in housing. Finally, the interaction
effects between homeowner propensity and the afford-
ability measures are also present, including in most of the
fixed effects models. Homeowners in more unaffordable
regions are even less positively inclined towards gov-
ernment intervention in housing. The right figure of
Figure 1 demonstrates that the relationship between net
affordability and support for government intervention in
housing is substantially stronger for people with a high
propensity to be homeowners than those with a low
propensity (the negative slope is statistically significant
for both groups but larger in the latter case).
Finally, Tables A1 and A2 in the Online Appendix
show that the negative effects of housing cost relative
Table 2 Housing policy preferences 2006 (individual controls not shown).
Random effects
(1) (2) (3) (4) (5)
Average cost Owner’s cost Renter’s cost Cost/gross inc Cost/net inc
Homeowner propensity 0.65
∗∗
(0.26) 0.61
∗∗
(0.26) 0.78
∗∗∗
(0.26) 1.05
∗∗∗
(0.29) 0.75
∗∗∗
(0.26)
Housing variable 0.84
∗∗∗
(0.03) 0.69
∗∗∗
(0.03) 0.80
∗∗∗
(0.04) 3.12
∗∗∗
(0.18) 2.19
∗∗∗
(0.11)
Fixed effects (1) (2) (3) (4) (5)
Average cost Owner’s cost Renter’s cost Cost/Gross inc Cost/Net inc
Homeowner propensity 0.76
∗∗∗
(0.26) 0.75
∗∗∗
(0.26) 0.75
∗∗∗
(0.26) 1.00
∗∗∗
(0.28) 0.75
∗∗∗
(0.26)
Housing variable 0.67
∗∗∗
0.52
∗∗
0.28
∗∗
0.12 1.54
∗∗
Random effects (0.16) (1) (0.22) (2) (0.13) (3) (1.25) (4) (0.75) (5)
Average cost Owner’s cost Renter’s cost Cost/gross inc Cost/net inc
Homeowner propensity 0.36 0.45 0.28 0.81
∗∗
0.52
∗
(0.28) (0.28) (0.29) (0.33) (0.30)
Housing variable 0.48
∗∗∗
0.50
∗∗∗
0.22 2.07
∗∗∗
1.41
∗∗∗
(0.14) (0.12) (0.15) (0.70) (0.46)
Homeowner X variable 0.49
∗∗∗
0.27
∗
0.81
∗∗∗
1.47 1.07
∗
(0.18) (0.16) (0.20) (0.95) (0.62)
Fixed effects (1) (2) (3) (4) (5)
Average cost Owner’s cost Renter’s cost Cost/gross inc Cost/net inc
Homeowner propensity 0.49
∗
0.56
∗∗
0.37 0.72
∗∗
0.45
(0.28) (0.27) (0.28) (0.32) (0.29)
Housing variable 0.32 0.29 0.17 1.25 0.47
(0.21) (0.25) (0.19) (1.45) (0.89)
Homeowner X variable 0.47
∗∗
(0.18) 0.31
∗∗
(0.16) 0.64
∗∗∗
(0.20) 1.72
∗
(0.91) 1.37
∗∗
(0.60)
N7941 7941 7941 7076 7941
Standard errors in parentheses.
*p<0.10, ∗∗p<0.05, ∗∗∗p<0.01.
604 Journal of European Social Policy 31(5)

Table 3. BSAS 2010: Housing construction and spending priorities.
(1) (2) (3) (4) (5) (6) (7) (8)
Build Build Build Build Spend Spend Spend Spend
Homeowner 0.349
∗∗∗
0.388 0.331
∗∗∗
1.994
∗∗∗
0.100
∗∗∗
0.063 0.100
∗∗∗
0.571
∗∗
(0.057) (0.300) (0.056) (0.703) (0.017) (0.095) (0.017) (0.232)
Average housing cost 0.013 0.665
∗
——0.205
∗∗∗
0.353
∗∗∗
——
(0.244) (0.365) ——(0.073) (0.133) ——
Homeowner X cost —1.179
∗∗
(0.473)
———0.260
∗
(0.151)
——
Unaffordability ——4.988
∗∗∗
(1.480)
11.250
∗∗∗
(2.383)
——1.275
∗∗∗
(0.443)
3.086
∗∗∗
(0.898)
Homeowner X
unaffordability
—— —9.976
∗∗∗
(3.007)
—— —2.883
∗∗∗
(1.003)
Household income 0.012
∗∗
0.011
∗∗
0.013
∗∗
0.013
∗∗
0.003
∗
0.003
∗
0.002
∗
0.002
∗
(0.005) (0.005) (0.005) (0.005) (0.001) (0.001) (0.001) (0.001)
Social class 0.013 0.014 0.013 0.015 0.006 0.006 0.005 0.006
(0.014) (0.014) (0.014) (0.014) (0.004) (0.004) (0.004) (0.004)
Education 0.008 0.008 0.010 0.010 0.002 0.002 0.002 0.002
(0.012) (0.012) (0.012) (0.012) (0.004) (0.004) (0.004) (0.004)
Unemployed 0.054 0.070 0.046 0.049 0.071
∗
0.074
∗
0.068
∗
0.069
∗
(0.109) (0.109) (0.110) (0.111) (0.039) (0.039) (0.039) (0.039)
Retired 0.044 0.038 0.046 0.048 0.007 0.006 0.009 0.009
(0.078) (0.078) (0.077) (0.077) (0.022) (0.022) (0.022) (0.022)
White 0.153 0.133 0.108 0.081 0.030 0.025 0.037 0.028
(0.095) (0.094) (0.092) (0.092) (0.029) (0.029) (0.029) (0.028)
Age 0.003 0.003 0.003 0.003 0.000 0.000 0.000 0.000
(0.002) (0.002) (0.002) (0.002) (0.001) (0.001) (0.001) (0.001)
Female 0.148
∗∗∗
0.147
∗∗∗
0.150
∗∗∗
0.145
∗∗∗
0.013 0.013 0.013 0.012
(0.045) (0.045) (0.045) (0.045) (0.013) (0.013) (0.013) (0.013)
Constant 3.671
∗∗∗
3.220
∗∗∗
2.465
∗∗∗
0.969 0.155
∗∗
0.057 0.014 0.446
∗∗
(0.233) (0.297) (0.392) (0.595) (0.070) (0.102) (0.117) (0.218)
N2532 2532 2532 2532 2584 2584 2584 2584
Standard errors in parentheses.
*p<0.10, ∗∗p<0.05, ∗∗∗p<0.01.
Ansell and Cansunar 605

to gross or net income at the regional level are robust
to other regional controls including regional income,
homeownership, income tax burden and housing
burden, all taken from the EU-SILC. Homeowner-
ship rates and average income correlate negatively
with redistribution/housing support. Income tax
burden correlates negatively in the random effects
models. In sum, we have seen that in regions of Europe
with more expensive housing, particularly as compared
to disposable income, support for redistribution and
government spending on housing is lower and that this
effect appears to be particularly concentrated among
homeowners. This suggests a potential polarization of
European politics as homeowners in expensive areas
pull away from those less fortunately situated in the
housing market.
Housing affordability in the United Kingdom
We now turn to examine the political implications of
housing affordability in a single country: the United
Kingdom. We choose to analyse the UK for two
reasons. First, the UK has both relatively high un-
affordability, with an average ratio of housing costs
to disposable income of around 24% in 2010, and
high regional variation in affordability (the second
highest in the EU-SILC data after Greece). This
means affordability is highly salient politically in the
UK due to its highlevel, but that there is also sufficient
variation within the country to provide empirical
traction for how local differences shape attitudes.
Second, the UK has excellent data availability on
house prices at the regional and local level, which we
can match to surveys and election results. Finally, as
Fuller (2019) convincingly argues, the commodifi-
cation of the UK housing market makes housing
inequality particularly salient in the British context.
We begin by looking once more at survey data to
examine how regional affordability affects policy
attitudes. We employ the British Social Attitudes
Survey from 2010 which has a unique set of ques-
tions asking people about their attitudes towards
housing policy. The BSAS also codes people by
which of 12 UK regions they live in, which we can
match to housing affordability data from the EU-
SILC. We then turn to political outcomes. Here we
are able to look at an even lower level of aggregation,
calculating housing affordability at the constituency
level and seeing how changes in affordability are
associated with changes in voting patterns along
Britain’s long-standing left–right dimension.
Housing affordability and public attitudes
in Britain
We saw in our analysis above of regional data in
Europe that housing unaffordability was associated
with a divergence between homeowners and renters
in their preferences over government responsibility
for decent housing. With the ISSP 2006 however we
had to estimate the propensity to own a house. By
contrast, the British Social Attitudes Survey (BSAS)
in 2010 has both information about an individuals’
housing tenure and their views about housing policy.
Accordingly, we have a more direct test of whether
regional unaffordability polarizes the housing policy
preferences of owners versus renters.
We draw two housing policy questions from the
BSAS. First, we use a question about housing sup-
ply: ‘Would you support or oppose more homes
being built in your local area?’, which is on a five-
point scale from strongly oppose to strongly support.
Second, we use responses to questions asking what
respondents’first and second priority for extra
government spending would be. We code this as one
if respondents mention housing as a first or second
priority and zero otherwise.
We run a series of linear regressions with these
two dependent variables and use as our core predictors
a dummy for homeownership and an indicator of re-
gional housing –in turn we use monthly average
housing costs and these costs as a proportion of
disposable income (net unaffordability).
13
In odd
numbered models we include these items sepa-
rately, whereas in even numbered models we in-
clude their interaction as well. We also include a
wide array of individual-level controls: household
income, social class, education, age and dummies
for being female, unemployed, retired or White
British in ethnicity.
Models 1 to 4 of Table 3 look at support for
building new houses locally. In Models 1 and 3 we
see a strong negative impact of being a homeowner
606 Journal of European Social Policy 31(5)

on support for building houses: an effect twice as
large as being female or moving from the 10th to 90th
percentile in income. Clearly, on average home-
owners are much less likely to support construction.
However, this effect is much more pronounced in
areas with more expensive housing, as seen in the
negative coefficients on the interaction of home-
ownership with the two regional housing variables.
Renters and owners have greatly diverging
preferences, where housing is expensive both ab-
solutely and relative to incomes. This can be seen
clearly in the left panel of Figure 2, which uses net
unaffordability as the regional context variable.
14
Where housing is relatively cheap (on the left of the
graph) homeowners and renters have statistically in-
distinguishable preferences, but as we move to the most
unaffordable areas (on the right) renters and owners are
almost a full point apart in their preferences.
Models 5 to 8 show a similar story when we turn
to placing housing as a priority for government
spending. Again, we see homeowners are 10% less
likely than renters to view housing as a priority. But
this is an average effect –once we take regional
variation in affordability into account, we see again
that renters and owners have similar preferences in
the more affordable areas and diverge substantially in
unaffordable areas –by up to 20% points, as can be
seen in the right panel of Figure 2.
In sum, the BSAS survey data shows that
homeowners are generally unsupportive of inter-
ventionist housing policy –as we saw in our
analysis of the ISSP –but that this relationship is
dependent on local affordability. Where housing is
affordable, renters and owners are not polarized in
their views. This divergence only emerges in ex-
pensive areas.
Figure 2. Net affordability and support for construction (left) and placing housing as a priority for government spending
(right).
Ansell and Cansunar 607

Housing affordability and elections: Evidence
from British general elections 2010–2019
How do changes in the preferences over policies
driven by raising housing costs translate into elec-
toral outcomes? We conclude by examining the
electoral consequences of housing costs and un-
affordability in the UK. We do so by considering the
proportion of votes cast at the electoral constituency
level for the Conservative Party in England and
Wales. As we outlined above, our expectation is that
housing unaffordability should be correlated with
growing support for the Conservative Party given its
positive consequences for the (homeowning) ma-
jority of voters.
Our dependent variable is (changes in) the vote
share of the Conservative Party at the parliamentary
constituency in the General Elections of 2010, 2015,
Table 4. Changes in Conservative vote share, income and wealth.
(1) (2) (3) (4) (5) (6) (7)
Dcons Dcons Dcons Dcons Dcons Dcons Dcons
Δmedian earnings 0.006 0.004 0.004 ————
(0.005) (0.004) (0.004) ————
Δearnings
inequality
4.291
∗∗
(1.889)
3.391
∗∗
(1.463)
3.248
∗∗
(1.365)
————
Δmedian house
price
—0.029
∗∗∗
(0.005)
— ————
Δin hp inequality —0.571
∗
(0.311)
— ————
Δ90p house price ——0.002
∗∗
(0.001)
————
Δ10p house price ——0.044
∗∗∗
(0.009)
————
Δhousing
affordability
———0.463
∗∗∗
(0.100)
———
Δp90 house costs —————0.143
∗∗∗
(0.019)
0.054
∗∗
(0.023)
Δp10 house costs ————0.860
∗∗∗
(0.234)
—0.701
∗∗
(0.244)
Δunemployment 1.801
∗∗∗
1.650
∗∗∗
1.735
∗∗∗
1.734
∗∗∗
1.777
∗∗∗
1.764
∗∗∗
1.768
∗∗∗
(0.499) (0.454) (0.447) (0.474) (0.468) (0.498) (0.470)
% owner 0.054
∗∗
0.068
∗∗∗
0.064
∗∗
0.059
∗∗
0.060
∗∗∗
0.050
∗∗
0.057
∗∗
(0.018) (0.020) (0.021) (0.018) (0.018) (0.021) (0.019)
% White 6.163
∗
8.239
∗∗
8.221
∗∗
7.082
∗∗
6.219
∗
8.934
∗∗
7.205
∗
(2.894) (3.292) (3.388) (3.018) (2.778) (3.722) (3.360)
% over 60 0.081 0.105 0.105 0.094 0.105 0.073 0.098
(0.065) (0.069) (0.068) (0.068) (0.072) (0.061) (0.069)
% non-UK 0.226
∗∗∗
0.266
∗∗∗
0.259
∗∗∗
0.246
∗∗∗
0.224
∗∗∗
0.277
∗∗∗
0.243
∗∗∗
(0.058) (0.060) (0.060) (0.052) (0.063) (0.043) (0.059)
Constant 14.521
∗∗∗
16.636
∗∗∗
16.326
∗∗∗
15.020
∗∗∗
14.079
∗∗∗
16.933
∗∗∗
15.047
∗∗∗
(3.883) (3.975) (4.126) (3.713) (3.950) (3.792) (4.155)
Year FE Yes Yes Yes Yes Yes Yes Yes
Region FE Yes Yes Yes Yes Yes Yes Yes
N1710 1710 1710 1712 1712 1712 1712
Standard errors in parentheses.
*p<0.10, ∗∗p<0.05, ∗∗∗p<0.01.
608 Journal of European Social Policy 31(5)

2017 and 2019. As economic controls, for each
electoral constituency and year, we use data from the
Office for National Statistics, including estimates of
the median gross weekly pay in each constituency,
information on gross income reported for each decile
and unemployment rates. Data on house prices
comes from the UK Land Registry and includes all
sales of houses and apartments in each constituency.
Finally, we obtain demographic data from the 2011
Census in England and Wales (ethnicity, age struc-
ture and immigration status).
We assess voter responses to changes housing
prices and (in)affordability using a first-difference
model. For constituency iin election year t, we es-
timate the following equation
ΔConservativeit ¼βΔAit þλXit þθtþγtþεit (1)
where our outcome variable is ΔConservative
it
,the
change in the Conservative Party’s vote share from
the preceding election at election year tin precinct i,
ΔA
it
is a vector of changes in the variables of in-
terest: median earnings, median house price level,
10th and 90th percentile house prices, earnings
inequality (operationalized as mean earnings/
median earnings), house price inequality (p90
house price/p10 house price), unemployment and
housing cost index [measured as median house
price/(weekly median earnings*52)]. We also in-
clude 10th percentile housing costs (measured as
10th percentile house price/(monthly median earn-
ings*52)) and 90th percentilehousing costs (measured
as 90th percentile house price/(weekly median earn-
ings*52)). X
it
is a vector of static control variables:
homeownership rate, percent white, share of residents
over the age 60 and share of the residents who were
born outside the UK, all taken from the 2011 UK
Census. θ
i
and γ
t
are region and election year fixed
effects, respectively. We also take account of auto-
correlation in the error term ε
it
by clustering standard
errors at the regional level.
We begin in Table 4 with the absolute effect of
house prices on Conservative support. Model 1 omits
house prices, focussing on the impact of local labour
market changes on the Conservative vote share. It
includes changes in median earnings, earnings in-
equality, unemployment levels, socio-economic
controls at the constituency level along with
dummies for the election years and 10 regions of
England and Wales. Changes in median earnings do
not significantly affect support for the Conservatives,
whereas increases in the levels of unemployment do
significantly increase support. Conversely, we see
that increases in earnings inequality significantly
decrease the Conservative vote share.
We now turn to the effect of house prices. Model 2
uses the change in median house prices and house
price inequality at the constituency level, along with
the changes in income and socio-economic controls,
election fixed effects and region fixed effects. The
results indicate that both changes in house prices and
house price inequality affect the Conservative vote
share significantly, albeit in different directions. We
see that median house price growth increases support
for the Conservatives. The estimated impact of a
£50,000 increase in median house prices is associ-
ated with a roughly 1.5 point increase in the Con-
servatives vote share. By contrast, increased
inequality in house prices, measured as the ratio of
90th percentile and 10th percentile house prices,
significantly decreases support for the Conservatives.
Models 3 regresses changes in the 10th and 90th
percentile of housing prices, on the changes in the
Conservative vote share. A £50,000 increase in the
90th percentile house price is associated with a
roughly 0.1 point increase in the Conservatives vote
share. Increases in the 10th percentile of house prices
have a much stronger positive impact on the electoral
performance of the Conservative party: a £50,000
increase is associated with a 2.2 point increase in the
Conservatives vote share. The coefficient on changes
in the 10th percentile of house prices is larger in
magnitude and statistically more significant than the
coefficient of 90th percentile of house prices. In other
words, prices going up in areas with less well-off
homeowners –traditionally Labour voters –appears
to convert them into Conservatives.
What about housing affordability –that is, the level
of house prices relative to income? Models 4 to 7
examine changes in housing affordability and con-
servative vote share, presenting a series of models
using housing costs and socio-economic character-
istics of constituencies. Model 4 shows that where
housing costs relative to income rose –that is,
Ansell and Cansunar 609

affordability declined –the Conservatives picked up
votes. Models 5–7 examine the relationship between
local changes in the affordability of expensive and cheap
houses in each constituency on the Conservative votes.
We find that both increases in expensive housing un-
affordability (measured as the ratio of 90th percentile
house prices to yearly median earnings) and cheap
housing affordability (measured as the ratio of 10th
percentile house prices to yearly median earnings) draw
more votes for the Conservatives.
Conclusion
In this article, we have explored how declining housing
affordability across Europe affects preferences for so-
cial policy and political outcomes. We have found
consistent evidence that declining affordability –driven
by increases in housing prices –decreases support for
redistribution, especially among homeowners, across
Europe and increases votes for the Conservative party
in the UK. Our analysis contributes to the literature that
highlights the centrality of wealth, especially home-
ownership, for social policy demand and political
outcomes (Fuller et al., 2020;Schwartz, 2009). We
theoretically argue and empirically show that relative
wealth is powerful in shaping demand for redistribution,
housing policy and support for conservative parties.
What are the broader theoretical lessons of these
findings? One is that house prices have a significant
impact on social policy preferences and voting be-
haviour, not only in absolute terms but also in relative
terms. We find that growing unaffordability creates
support for less government intervention in labour
market inequality or housing and, consequently more
support for the political right. We argued that this
result is driven by homeowners, seeking to protect
their gains of appreciation from being taxed away or
undermined by growing housing supply. We show
that this is not simply a function of places with higher
incomes having more expensive property –the ratio
of housing costs to incomes is a core factor driving
this shift in preferences.
Another lesson concerns the consequences of
wealth inequality for political polarization. Even if on
average, rising affordability reduces the demand for
intervention, that may not be the case for renters, locked
out of booming markets. Rising unaffordability drives a
political wedge between owners and renters and may be
an important contributor to the growth of political
polarization, including that attributed to age (McCarty
et al., 2016;O’Grady, 2019). As Dewilde and Flynn
(2021) show in this special issue, homeownership is
particularly unequally distributed among the young,
potentially detaching many younger Europeans from
the housing market permanently, with major economic
and political consequences. The political relevance of
this type of polarization will of course depend on the
ability of actors defined by their wealth to coalesce
around housing policy –if income or labour market risk
cutacrosstheinterestsofhomeownersthismightre-
duce the size of the ‘anti-redistribution’coalition.
Cultural divides might also split homeowners from one
another, as in the case of the Brexit divide noted by
Adler and Ansell (2020), which has produced the
unusual example of a Conservative Party toying with
easing planning laws.
We conclude by noting the policy consequences
of our argument and findings. Presuming that centre-
right parties are inherently cautious about market
interventions, particularly in residential housing, we
should not expect that the aggregate growth in
housing unaffordability will meet with a thermostatic
political response. A number of articles in this vol-
ume have suggested an array of redistributive poli-
cies from rental controls to tax reform to help
younger non-owners (Dewilde and Flynn, 2021;
Nolan et al., 2021). Instead, in the medium run,
growing housing unaffordability is likely to lead to
greater success for right-wing parties and less interest
in redistribution, thereby underpinning the overall
trend of rising property prices. Accordingly, the
surge in house prices we have seen across Europe is
self-reinforcing. The beneficiaries of unaffordability
will prefer to keep policies and parties in place that
keep prices high and rising. However, as we have
seen, they do so at the cost of growing polarization
between renters and owners.
Acknowledgements
The authors gratefully acknowledge the helpful criticisms
and suggestions provided by two anonymous reviewers,
the super-reviewer, the special issue editors Brian Nolan
and Ive Marx, and the editors of the Journal of European
610 Journal of European Social Policy 31(5)

Social Policy. We thank Mads Elkjaer, Laure Bokobza and
Jacob Nyrup for their comprehensive feedback.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest
with respect to the research, authorship, and/or publication
of this article.
Funding
This project has received funding from the European
Research Council (ERC) under the European Union’s
Horizon 2020 research and innovation programme, grant
agreement number 724949. The ERC project code for this
project is WEALTHPOL.
ORCID iD
Ben Ansell https://orcid.org/0000-0001-8371-0507
Supplemental material
Supplemental material for this article is available online.
Notes
1. Here, we follow the standard Meltzer and Richard
(1981) view that individuals with higher incomes or
wealth will be more concerned about taxation and
benefit relatively less from receiving transfers. This is
not to foreclose the possibility, noted by Shayo (2009),
Rueda (2018) and others that non-material motivations
may also drive redistributive preferences. However,
Ansell (2014) suggests house prices matter for re-
distributive preferences net of ideology, race and re-
ligiosity, among other non-material factors.
2. Exactly where new houses are built may alter pref-
erences. Hankinson (2018) shows that San Francisco
renters often oppose new construction locally even
when they support it nationally. Our empirical anal-
ysis, however, looks at national housing policy, where
we expect renters to have clearer attitudes.
3. This article is based on data from Eurostat, EU-SILC
2006–2017, used under agreement RPP 346/2018-
ECHP-EU-SILC. The responsibility for all conclusions
drawn from the data lies entirely with the authors.
4. Unfortunately, countries vary substantially in their
regional coverage in the EU-SILC. Denmark, for
example, lacks any regional information. Others vary
in which years they include regions and, on occasion,
which regions they include. In a few cases, we have
had to match the data to a slightly different year: for the
ISSP 2009, we use 2006 data for Germany, 2007 for
Switzerland and 2010 for the United Kingdom. In-
cluding country-fixed effects in some of our analyses
means that we focus solely on regional variation in
affordability in those cases. We also replicate our
analysis on solely those countries with multiple re-
gions in the Appendix in Tables A2 and A3.
5. We use only those groupings with 50 or more
households to conform with EU-SILC requirements.
6. This figure includes Ireland, Italy, the Netherlands and
Portugal, which are not in our analyses as they lack
data on key individual-level variables.
7. In the Appendix Tables A4 and A5, we also examine
models that include household size –results are ex-
tremely similar to those in the main text.
8. The variable used by the ISSP asks people ‘about how
much money would be left if the home or apartment
you and/or your immediate family live in was sold, and
any debts on it, such as a mortgage or personal loan,
would have been paid off’. From this, we obtain a
series of estimates among homeowners of their per-
ceived equity in the house. Our interest is in comparing
that group who own to those who answer the question
that they do not own a home. There is some miss-
ingness in answers to this question –of a possible
14,486 people who were asked this question, 2879 do
not provide an answer –but that group’s income and
age appear to be intermediate compared to home-
owners and renters –they do not appear to be sys-
tematically different. Portugal is the one country with
very high missingness on this question (68%) so we
omit it from the analyses.
9. Comparing the actual rate of homeownership from the
EU-SILC data to that in the responses in the ISSP
2009, we find that the survey on average contains more
homeowners than the national statistics but that by
country this amount varies from the official figure by
between 3 and 11%.
10. Appendix Table A2 removes countries without re-
gional variation in housing affordability –results are
very similar to those in Table 1.
11. In the country-fixed effects models, this is only statisti-
cally significant for the housing to net income variable,
which suggests that country-level affordability effects are
Ansell and Cansunar 611

likely to be more important in shaping how homeowners
react than regional differences net of country averages.
12. Appendix Table A3 removes countries without regional
variation in housing affordability –here we see somewhat
weaker effects of the homeownership interaction vari-
able, though the direct variable remains robust
13. Results are similar if we use a logit specification –see
Figure A4. We take these measures as before from the
EU-SILC household data at the UK regional level –we
are able to use data for 12 British regions (region is the
lowest level of aggregation in the BSAS, 2010). We
use the linear probability model for easy interpret-
ability of coefficients
14. Figure A3 in the Appendix shows the predicted
support with average cost as the context variable.
References
Adler, D. and Ansell, B. (2020) ‘Housing and populism’,
West European Politics 43(2): 344–365.
Albouy, D., Ehrlich, G. and Liu, Y. (2016) Housing De-
mand, Cost-Of-Living Inequality, and the Afford-
ability Crisis. Technical report. National Bureau of
Economic Research, Cambridge MA.
Ansell, B. (2014) ‘The political economy of ownership:
housing markets and the welfare state’,American
Political Science Review 108(2): 383–402.
Ansell, B. and Adler, D. (2019) ‘Brexit and the politics of
housing in Britain’,The Political Quarterly 90:
105–116.
Ansell, B.W., Broz, J.L. and Flaherty, T. (2018) ‘Global
capital markets, housing prices, and partisan fiscal
policies’,Economics & Politics 30(3): 307–339.
Brady, D. and Bostic, A. (2015) ‘Paradoxes of social
policy: welfare transfers, relative poverty, and re-
distribution preferences’,American Sociological
Review 80(2): 268–298.
Cansunar, A. (2021) ‘Who is high-income, anyway? Social
comparison, subjective group-identification, and
preferences over progressive taxation’,Journal of
Politics 112: 14–27. DOI: 10.1086/711627
Castles, F. (1998) ‘The really big trade-off: home own-
ership and the welfare state in the new world and the
old’,Acta Politica 33: 5–19.
Conley, D. and Gifford, B. (2006) ‘Home ownership, social
insurance, and the welfare state’,Sociological Forum
21(1): 55–82.
Cusack, T., Iversen, T. and Rehm, P. (2006) ‘Risks at work:
the demand and supply sides of government redistribu-
tion’,Oxford Review of Economic Policy 22(3): 365–389.
Dewilde, C. (2018) ‘Explaining the declined affordability
of housing for low-income private renters across
Western Europe’,Urban Studies 55(12): 2618–2639.
Dewilde, C. and Flynn, L. (2021) ‘Post-crisis develop-
ments in young adults’housing wealth’,Journal of
European Social Policy 14(1): 121–127.
Dustmann, C., Fitzenberger, B. and Zimmermann, M.
(2018) Housing Expenditures and Income Inequality,
ZEW-Centre for European Economic Research.
Discussion Paper (18-048), Leibniz.
Fuller, G.W. (2019) The Political Economy of Housing
Financialization. Newcastle-upon-Tyne, UK: Agenda
Publishing.
Fuller, G.W., Johnston, A. and Regan, A. (2020) ‘Housing
prices and wealth inequality in western Europe’,West
European Politics 43(2): 297–320.
Hankinson, M. (2018) ‘When do renters behave like home-
owners? High rent, price anxiety, and NIMBYism’,
American Political Science Review 112(3): 473–493.
Kemeny, J. (1981) Myth of Home-Ownership. London:
Routledge.
Knoll, K., Schularick, M. and Steger, T. (2017) ‘No price
like home: global house prices, 1870–2012’,Amer-
ican Economic Review 107(2): 331–353.
McCarty, N., Poole, K.T. and Rosenthal, H. (2016) Po-
larized America: The Dance of Ideology and Unequal
Riches. Cambridge, MA: MIT Press.
Meltzer, A.H. and Richard, S.F. (1981) ‘A rational theory
of the size of government’,The Journal of Political
Economy 89(5): 914–927.
Nolan, B., Palomino, J.C., van Kerm, P., et al (2021)
‘Inheritance, gifts and the accumulation of wealth for
low-income households’,Journal of European Social
Policy 12: 312–353.
O’Grady, T. (2019) ‘Age cleavages across European
countries and issue dimensions since the 1980s’,
Wor kin g Pa per, https://files.osf.io/v1/resources/r8xbg/
providers/osfstorage/5d6cd9c5536cf5001786d11b
Rehm, P. (2011) ‘Social policy by popular demand’,World
Politics 63(2): 271–299.
Rueda, D. (2018) ‘Food comes first, then morals: redis-
tribution preferences, parochial altruism, and immi-
gration in Western Europe’,The Journal of Politics
80(1): 225–239.
612 Journal of European Social Policy 31(5)

Scheve, K. and Slaughter, M. (2001) ‘What determines
individual trade-policy preferences?’,Journal of In-
ternational Economics 54: 267–292.
Schwartz, H. (2009) Subprime Nation: American Power,
Global Capital, and the Housing Bubble. Ithaca, NY:
Cornell University Press.
Shayo, M. (2009) ‘A model of social identity with an
application to political economy: nation, class, and
redistribution’,American Political Science Review
103(2): 147–174.
Stegmueller, D. (2013) ‘Modeling dynamic preferences: A
Bayesian robust dynamic latent ordered probit
model’,Political Analysis 21(3): 314–333.
Svallfors, S. (1997) ‘Worlds of welfare and attitudes to
redistribution: a comparison of eight western nations’,
European Sociological Review 13(3): 283–304.
Ansell and Cansunar 613