Health Effects of Housing Insecurity and Unaffordability in the General Population in Barcelona, Spain

Abstract

While the negative effects of housing insecurity and unaffordability on health are well known, most of the studies in Spain have focused on very specific social groups so their findings cannot be extrapolated to the general population. The aim of this study is to assess the effects of housing stress and risk of displacement due to economic reasons, and their combined effect, on the mental and physical health of the general population from a middle-income neighborhood of Barcelona. We conducted a cross-sectional study using a household health survey which included respondents from a representative sample of 1202 non-institutionalized residents (> 18 years old) of the Horta neighborhood. We carried out a descriptive analysis, estimated the prevalence of poor mental and self-rated health (stratifying by the independent variables) and fitted robust Poisson regression models to estimate the effects of housing stress (HS) and the risk of forced displacement (RD) on self-rated health and mental health (GHQ-12). All analyses were stratified by sex. We found a higher likelihood of poor general and mental health among people affected by HS and/or RD compared to those not affected by HS and/or RD. A graded effect of HS and RD emerged mainly on mental health, even after adjusting by socioeconomic variables and housing tenure. The serious problem of housing insecurity and unaffordability in Spain is a widespread public health issue. Evidence-based public policies to improve well-being and health of people under this threat are urgently needed.

Introduction

For many years, housing has being recognized as an important social determinant that can influence health through both physical and socioeconomic dimensions [1–3]. The latter dimension is related to several issues such as housing affordability and secure tenure, which allow people to have a stable home and a place that protects privacy, contributes to physical and mental health, and supports the development of a life plan and social integration of its inhabitants [4]. Housing affordability issues encompass financial stress due to housing costs (rent and mortgage-related costs and the overall costs of housing consumption, including utility bills)—known as housing stress—and the quality and location trade-offs made by people when choosing a house (e.g., being willing to pay a little more than planned to live in a neighborhood with better access to goods and services) [5]. When housing affordability problems worsen, there is a risk of forced displacement due to more severe housing insecurity situations such as foreclosure and eviction or illegal housing tenure (e.g., squatting) [6, 7].

Several studies have reported the negative consequences of housing unaffordability and housing insecurity on health [8–17]. The effects on mental health are fairly clear. Persons having trouble paying their mortgage have higher rates of psychological distress and depression, which worsen after the start of the foreclosure process [17]. An association has also been found between self-reported unaffordable housing and poor psychological health [14], and between moderate housing strain and increased psychological distress [9]. In contrast, the findings on physical health are less clear. A systematic review on the topic reported an association between poor self-rated health and unaffordable housing; however, no relationship between somatic health and housing distress was observed in studies including health status in a prior period [10]. Regarding more severe housing insecurity situations such as the threat of eviction, a recent systematic review reported negative effects on both mental (e.g., depression, anxiety, psychological distress, and suicide) and physical health (poor self-reported health, high blood pressure, and child maltreatment) [16].

Although housing unaffordability is a widespread issue, it strongly affects places where housing has been highly commodified [18, 19]. For example, Spain has a basically market-driven housing system with less than 2% of public social housing. Several policies implemented from the second half of last century (mainly at the national level) have led housing to become in an investment and tool for economic growth rather than a social right [20, 21]. This situation has worsened since the onset of the subprime crisis in 2007. Thus, the prevalence of having to leave a home within a 6-month period due to economic issues was 7.1% in 2012, while 752,663 foreclosure were initiated and 657,070 evictions ordered between 2008 and 2018 [22]. Barcelona is the city most affected by these factors and is also undergoing an intense process of gentrification and a sharp rise in rents, making access to adequate housing even more difficult [23, 24].

Some studies have reported the negative effects of unaffordability and housing insecurity on mental and physical health in Spain (and Barcelona) [12, 15, 25, 26]. However, those studies focused on highly specific groups (i.e., activists from social movements for affordable housing and disadvantaged people served by charitable institutions), and consequently, their findings cannot be extrapolated to the general population. The aim of this study was to assess the effects of subjective housing stress and risk of displacement for economic reasons, as well as their combined effect, on the mental and physical health of the general population living in a middle-income neighborhood of Barcelona.

Methods

Study Design, Population, and Information Sources

This cross-sectional study used the “Salut als Carrers” household health survey as the data source. “Salut als Carrers” is a project aiming to evaluate the effects on health of an urban renewal plan promoted by the City Council of Barcelona [27]. This survey is the baseline of this evaluation in Horta, a neighborhood in Barcelona. The survey was performed in a representative sample of non-institutionalized residents. In this study, we included only residents aged ≥ 18 years, with a sample size of 1202 individuals (601 women) [28]. The fieldwork to collect the information was done from May to September, 2018. The survey used a stratified random sampling of the neighborhood’s population by age and gender strata (200 individuals in each strata) with replacement of non-responding sampled individuals.

Description of the Variables

Dependent Variables

Our dependent variables were self-rated health and mental health. A five-level item enquired about self-rated health through the question: “In general, how would you rate your health today?” with the possible choices being (1) “very good,” (2) “good,” (3) “moderate,” (4) “bad,” or (5) “very bad” [29, 30]. Self-rated health was dichotomized into good self-rated health (“very good,” “good,” or “moderate”), and poor self-rated health (“bad” or “very bad”). Mental health was evaluated using the 12-item version of the General Health Questionnaire (12-GHQ), and participants scoring ≥ 3 were classified as having a high probability of poor mental health [31].

Independent variables

Housing stress (HS) was measured based on an adaptation of the Spanish Living Conditions Survey 2018 question [32]: “Considering the total cost of housing (including rent or mortgage, taxes, heating, electricity and other housing-related costs), how do you meet this expenditure?” “Very easily,” “easily,” “fairly easily,” “with some difficulty,” “with difficulty,” “with great difficulty.” Then, HS was dichotomized into people with HS (those clearly reporting stress: “with difficulty,” “with great difficulty”) and without HS (the remaining categories).

Risk of forced displacement (RD) for financial reasons was measured based on the question from the European Quality of Life Survey 2016 [33]: “How likely or unlikely do you think it is that you will need to leave your accommodation within the next 6 months because you can no longer afford it?” “Very likely,” “rather likely,” “neither likely nor unlikely,” “rather unlikely,” “very unlikely.”. RD was dichotomized in affected people (“very likely” and “rather likely”) and unaffected people (“neither likely nor unlikely,” “rather unlikely,” and “very unlikely”).

Finally, using both dichotomized variables, we created HS+RD to estimate the combined effect, which included three categories: people not affected by either HS or RD, people under HS without RD, and people under HS with RD. Only three individuals reported being under RD without HS, and were considered as missing values in all the analyses.

Adjustment variables

The adjustment variables were age, nationality (European countries and non-European countries), occupational social class in three categories from the classification of the Spanish Society of Epidemiology (where class I is the highest social class and class III the lowest) [34], employment status and type of housing tenure (“owner-occupancy,” “mortgage,” “tenancy,” “public social housing,” and “others”).

Statistical analysis

All analyses were stratified by sex and the sample was weighted (using the sample design) [28]. We performed a descriptive analysis of all variables reporting absolute and relative frequencies (Table 1).

Table 1.

Descriptive sociodemographic, independent variables, and health outcomes

	Women		Men
	n	%	n	%
Mental health
Good	493	75.9	478	86.5
Poor	155	23.9	74	13.3
Missing	1	0.2	1	0.2
Total	649	100.0	553	100.0
Self-rated health
Good	465	71.7	442	80.0
Poor	184	28.3	110	20.0
Missing	--	--	--	--
Total	649	100.0	553	100.0
Age
18–44	231	35.6	228	41.3
45–64	199	30.7	183	33.0
> 64	219	33.7	142	25.7
Missing	--	--	--	--
Total	649	100.0	553	100.0
Nationality
EU	561	86.4	481	87.0
Non-EU	88	13.6	72	13.0
Missing	--	--	--	--
Total	649	100.0	553	100.0
Social class
Class I	188	29.0	134	24.2
Class II	167	25.7	152	27.5
Class III	275	42.4	259	46.8
Missing	19	2.9	8	1.5
Total	649	100.0	553	100.0
Employment status
Paid workers	310	47.7	305	55.1
Unemployment	43	6.6	45	8.2
Others	297	45.7	203	36.7
Missing	--	--	--	--
Total	649	100.0	553	100.0
Homeownership
Tenure	312	48.1	253	45.8
Mortgage	141	21.7	110	19.9
Renting	167	25.7	166	30.0
Public social housing	6	0.9	6	1.0
Others	16	2.4	12	2.1
Missing	8	1.3	7	1.2
Total	649	100.0	553	100.0
Housing stress
No	498	76.8	441	79.8
Yes	126	19.4	83	15.1
Missing	25	3.9	28	5.1
Total	649	100.0	553	100.0
Risk of forced displacement
No	617	95.0	537	97.1
Yes	20	3.2	14	2.5
Missing	12	1.8	2	0.4
Total	649	100.0	553	100.0
Housing stress with risk
Not affected	485	74.8	434	78.5
Housing stress without risk	108	16.6	75	13.5
Housing stress with risk	17	2.6	8	1.4
Missing	39	6.0	37	6.6
Total	649	100	553	100

All analysis were weighted for the Horta neighborhood level

EU European Union

Then, we estimated the prevalence of poor mental and self-rated health for all independent variables. Next, we fitted a set of robust Poisson regression models to compute prevalence ratios of poor mental and self-rated health, and their 95% confidence intervals (95%CI), for all independent variables, firstly only age-adjusted (PR_a) and then adjusted by age, socioeconomic variables (origin, social class, employment status) and type of housing tenure (PR_b) (Table 2). All data analysis was conducted separately for men and women using Stata version 15.

Table 2.

Relative frequency of housing stress, risk of displacement and housing stress + risk of displacement by sociodemographic variables

	Housing stress				Risk of displacement				HS+RD
	Women		Men		Women		Men		Women			Men
	(%)	p-value	(%)	p-value	(%)	p-value	(%)	p-value	HS (%)	HS+RD (%)	p-value	HS (%)	HS+RD (%)	p-value
Age
18-44	25.0	0.10	16.4	0.59	6.8	<0.01	3.5	0.38	0.20	5.6	0.02	14.6	2.2	0.49
45-64	17.4		17.2		2.5		2.0		15.6	2.1		16.5	0.5
>64	17.7		13.3		3.2		1.5		17.1	0.5		15.4	1.6
Nationality
EU	18.6	0.02	14.7	0.06	2.5	0.01	0.9	<0.01	16.6	2.2	0.01	13.8	0.7	<0.01
Non-EU	30.8		23.6		7.8		13.3		24.7	6.9		19.1	6.7
Social class
Class I	16.1	0.04	11.3	0.12	6.0	0.03	2.1	0.49	12.2	4.5	0.02	11.0	0.6	0.23
Class II	16.5		14.2		0.7		1.5		15.8	0.8		13.1	0.8
Class III	25.0		19.0		2.6		3.4		23.1	2.6		16.9	2.3
Employment status
Paid workers	21.4	0.23	15.5	0.05	4.9	0.04	2.8	0.44	17.8	4.0	0.20	14.7	1.1	0.27
Unemployment	28.6		28.4		5.3		4.7		27.7	2.5		24.6	2.9
Others	17.7		13.7		1.2		1.6		16.1	1.6		12.0	1.7
Homeownership
Tenure	13.3	<0.01	9.3	<0.01	0.0	<0.01	0.0	<0.01	12.9	0.4	<0.01	9.3	0.0	<0.01
Mortgage	18.5		16.0		0.7		0.0		17.8	0.7		16.0	0.0
Renting	34.3		25.4		11.1		8.4		26.9	8.6		21.3	4.9
Others*	20.4		11.8		5.0		0.0		15.8	5.3		11.8	0.0

*For this analysis public social housing is included in “others”; HS Housing stress without risk of displacement, HS+RD: housing stress and risk of displacement

Results

Descriptive Analysis

Analysis of health status showed that 28.3% of women and 20% of men reported poor self-rated health, and 23.9% and 13.3% reported poor mental health, respectively (Table 1). Regarding age and gender, the distribution was similar between categories due to the sample design. Most women and men were from EU countries (86.4% and 87%, respectively), belonged to social class III (manual workers) (42.4% and 46.8%, respectively) and reported being in paid work at the interview (47.7% and 55.1%, respectively). Regarding housing tenure, most people lived in their own dwellings (totally paid or with a mortgage) (69.8% of women and 65.7% of men), followed by renters (25.7% and 30%, respectively). Only 0.9% of women and 1% of men reported living in public social housing (Table 1). Among women, 19.4% reported being under HS and 3.2% under RD, while among men these percentages were 15.1% and 2.5%, respectively (Table 1). Only 2.6% of the women and 1.4% of men reported HS and RD jointly.

The groups most affected by HS were women, people born in non-EU countries, those in class III (among women), unemployed persons (among men) and renters. RD mostly affected people born in non-EU countries and renters (Tables 1 and 2).

The Effects of Housing Stress and Risk of Forced Displacement, and Their Combined Effect, on Self-rated and Mental health

The prevalence and prevalence ratios of poor self-rated health and poor mental health among people affected and not affected by HS, RD and their combined effect are shown in Table 3.

Table 3.

Prevalence and prevalence ratios of poor self-reported health and poor mental health among both people affected and not affected by HS, RD, and their combined effect by sex

	Poor self-reported health						Poor mental health
	%	p value	PRa	(95% CI)	PRb	(95% CI)	%	p value	PRa	(95% CI)	PRb	(95% CI)
Women
Housing stress
No	25.3		1		1		19.2		1		1
Yes	45.0	< 0.001	1.95	1.55–2.45	1.83	1.44–2.33	47.0	< 0.001	2.46	1.87–3.23	2.22	1.66–2.96
Risk of displacement
No	27.9		1		1		22.6		1		1
Yes	27.1	0.94	1.82	0.86–3.88	1.54	0.65–3.67	44.7	0.03	1.96	1.12–3.41	1.54	0.88–2.69
HS+RD
Not affected	25.2		1		1		18.6		1		1
HS without RD	45.3		1.89	1.49–2.40	1.8	1.40–2.31	44.4		2.38	1.77–3.20	2.17	1.59–2.96
HS with RD	39.0	< 0.001	2.55	1.40–4.63	2.25	1.18–4.28	60.2	< 0.001	3.23	2.03–5.14	2.88	1.87–4.45
Men
Housing stress
No	18.4		1		1		10.9		1		1
Yes	31.0	< 0.001	1.82	1.33–2.50	1.61	1.16–2.24	27.7	< 0.001	2.57	1.68–3.93	2.18	1.39–3.42
Risk of displacement
No	19.8		1		1		12.7		1		1
Yes	23.4	0.53	1.74	0.81–3.73	1.43	0.60–3.39	35.0	0.028	2.89	1.34–6.20	2.84	1.30–6.21
HS+RD
Not affected	18.3		1		1		10.8		1		1
HS without RD	31.5		1.83	1.32–2.55	1.64	1.17–2.32	24.4		2.27	1.45–3.54	1.96	1.20–3.21
HS with RD	18.4	0.008	1.33	0.52–3.44	1.09	0.43–2.77	51.4	< 0.001	4.99	2.32–10.78	4.60	2.05–10.28

PR_a age-adjusted, PR_b adjusted by socioeconomic variables and type of tenure

Housing Stress

The prevalence of poor self-rated health was 45% among women and 31% among men affected by HS, compared with 25.3% and 18.4% among those not affected, corresponding to a PR_a (95%CI) of 1.95 (1.55–2.45) and 1.82 (1.33–2.50), respectively. The prevalence of poor mental health was 47% among women and 27.7% among men affected by HS, compared with 19.2% and 10.9% among those not affected, corresponding to a PR_a (95%CI) of 2.46 (1.87–3.23) and 2.57 (1.68–3.93), respectively. These associations remained significant after adjustment by socioeconomic variables and type of tenure.

Risk of Forced Displacement

The prevalence of poor self-rated health was 27.1% among women and 23.4% among men affected by RD, compared with 27.9% and 19.8% among those not affected, corresponding to a PR_a (95%CI) of 1.82 (0.86–3.88) and 1.74 (0.81–3.73), respectively. The prevalence of RD was 44.7% among women and 35% among men affected by RD, compared with 22.6% and 12.7% among those not affected, corresponding to a PR_a (95%CI) of 1.96 (1.12–3.41) and 2.89 (1.34–6.20), respectively. After the inclusion of adjustment variables, only the effect on mental health among men remained significant.

Combined Effect of HS and RD

The prevalence of poor self-rated health among women was 25.2% in those not affected by HS and RD, rising to 45.3% in those under HS without RD and 39% in those under HS with RD (PR_a 1.89 [1.49–2.40] and 2.55 [1.40–4.63], respectively). Among unaffected men, the prevalence was 18.3%, rising to 31.5% in those under HS without RD and 18.4% in those under HS with RD (PR_a 1.83 [1.32–2.55] and 1.33 [0.52–3.44], respectively).

A graded effect was observed for poor mental health among both women and men (Table 3). The prevalence of poor mental health among unaffected women was 18.6%, rising to 44.4% in those under HS without RD and 60.2% in those under HS with RD (PR_a 2.38 [1.77–3.20] and 3.23 [2.03–5.14], respectively). Among unaffected men, the prevalence was 10.8%, rising to 24.4% in those under HS without RD and 51.4% in those under HS with RD (PR_a 2.27 [1.45–3.54] and 4.99 [2.32–10.78], respectively).

Discussion

This is the first Spanish study to show the higher likelihood of poor general and mental health among people affected by HS and/or RD compared with unaffected persons in a representative sample of the general population and reveals a graded effect when exposures are added, mainly on mental health, even after adjustment by socioeconomic variables and housing tenure.

These findings are supported by prior local and international evidence on the negative effects of housing unaffordability and housing insecurity on health [10, 12, 15, 16, 35, 36]. For instance, three systematic reviews published in recent years concur that housing unaffordability, foreclosures, and the eviction process (a severe kind of RD) can have negative consequences on both mental health (depression, anxiety, and suicide) and physical health (poor self-rated health, high blood pressure, child maltreatment) [10, 16, 36]. At the local level, worse mental health (GHQ-12) has been reported among people affected by housing unaffordability attended by Cáritas Barcelona (a charity) compared with the lower class population of the city in 2015 [12]. Another study showed that the prevalence of poor mental and self-rated health was much higher in members of the Platform of People Affected by Mortgages of Catalonia (a social movement that fights for the right to housing) who were experiencing housing insecurity than in the general population [15]. However, both studies analyzed the health status of people from highly specific groups, hampering extrapolation of these findings to the general population.

Increased HS can lead to forced displacement, worsening the health consequences among affected individuals. Indeed, our findings are consistent and show a graded effect among women and men. Thus, people facing RD in addition to HS reported worse health than people only under HS and, in turn, those under HS reported worse health than unaffected individuals. This finding was clearer for mental health, because it is a sensitive indicator, probably being the first dimension of health status to be affected by legal and economic housing-related issues such as HS and RD [15, 37]. This is not surprising because these exposures are experienced by individuals as a personal failure and as a concealable stigma, and lead to feelings of insecurity, uncertainty, fear, embarrassment, isolation, and a lack of control over key aspects of daily living, all of them psychological changes that can lead to mental health problems [16, 26, 37–39].

This study, as well as previous studies on housing insecurity and unaffordability [12, 15, 25], have reported a higher prevalence of poor mental and self-rated health among women—whether affected by housing issues or not—than men, which is consistent with a large body of evidence on gender inequalities in health [12, 25, 38]. In addition, we report a higher prevalence of HS and RD in women than in men. This higher exposure to housing issues is supported by other authors such as Dymski et al who discussed how financial institutions targeted the subprime mortgages market to women, especially those from ethnic minorities and with lower socioeconomic status, which finally led to a higher proportion of women being affected by foreclosure and eviction in the USA [40, 41]. However, we should be cautious about inequalities on the effects of HS and RD on health, because, although we found some differences between women and men, we did not carry out statistical tests to assess significance, and prior evidence is unclear [15, 16, 35]. Further studies focused on the analysis of gender inequalities in the effects of housing insecurity and unaffordability on health are needed.

In addition to the aforementioned difference in exposures between women and men, we found other social groups that were most affected by HS and RD, such as younger people, those from the lowest social class, unemployed persons and people from non-EU countries. These findings are consistent with social health inequalities theory, which holds that one of the mechanisms explaining these inequalities is higher exposure to living conditions that may jeopardize health [3]. Other studies on the effects of housing insecurity and unaffordability on health have found similar results, reporting more exposure among persons with less education, ethnic minorities, people with lower incomes and unemployed people [9, 15, 42–44]. The distribution of HS and RD also varies according to type of housing tenure. We found a higher prevalence of both housing issues among renters than in persons with other kinds of tenure, coinciding with the change in the pattern of housing insecurity problems by type of tenure. The outbreak of the “subprime” crisis in 2008 and the international credit crunch first affected people with mortgages, leading to a large increase in foreclosures and evictions [22]. However, several policies to reactivate the real estate market focused on the rental sector instead of homeownership during the first half of the past decade have turned the former into an even more precarious and insecure type of tenure (e.g., shortening of the length of rental agreements, providing tax benefits to vulture funds, fostering luxury and holiday rental apartment market, etc.). This phenomenon is even more harmful given the insignificant public social housing stock (less than 2%) in Spain, which was also reflected in our findings [45]. All of these factors have contributed to a rapid increase in rental prices and a gentrification process (mainly in large cities such as Barcelona), which, in turn, have led to a severe scenario of housing insecurity and unaffordability among renters. For instance, 62.5% of evictions during 2018 were due to rent arrears [22].

This study has some limitations. Firstly, the sample may not have sufficient power to detect significant differences between some categories (e.g., RD), and therefore, our findings should be interpreted with caution. Secondly, both the dependent and independent variables are subjective variables, and consequently, individuals may have a poor perception of their residential situation, as well as a poor perception of their health. However, GHQ-12 is a well proven tool with psychometric characteristics that make it a reliable indicator of mental health [31, 46], while self-rated health has been reported as an accurate proxy of the burden of disease [29, 30, 47]. Despite these limitations, this study provides new data on the association between housing insecurity, unaffordability and health in the general population, a distinct context to that of most previous studies in Spain, which have focused on highly specific groups [12, 15, 25, 35].

Conclusion

The serious problem of housing insecurity and unaffordability in Spain is a public health issue which, unlike the first few years of the economic crisis when most affected individuals were homeowners with mortgages, is currently most frequent among renters than among persons with other types of tenure. This phenomenon has negative consequences on the health status of affected individuals and might increase social and health inequalities. Further research on this topic among the general population is needed to design evidence-based public policies to improve the well-being and health of people living under this threat.