Housing status, mortgage debt and financial burden as barriers to health among older adults in the U.S

Abstract

We examine relations between housing status, mortgage, financial burden, and healthy aging among older U.S. adults. We combine cross-sectional data from 2012 to 2014 Health and Retirement Study cohorts. Using regression models, we examined associations between owners and renters, mortgage and non-mortgage holders, financial strain, and difficulty paying bills, and poor self-rated health (SRH), heart condition (HC) and hospitalization (past two years). We find that compared to owners, renters had greater likelihood of poor SRH and hospitalization. Regardless of tenure, financial strain was associated with greater likelihood of poor SRH, HC and hospitalization, while difficulty paying bills was associated with poor SRH and HC. Mortgage holders had lower likelihood of poor SRH. Accounting for mortgage status, financial strain was associated with greater likelihood of poor SRH, HC and hospitalization, while difficulty paying bills was associated with poor SRH and HC. Associations between tenure or mortgage status and health were not modified by either financial burden factors. We conclude that there need to be more robust and inclusive programs that assist older populations with housing could improve self-rated health, with particular attention to renters, mortgage holders and those experiencing financial burden.

BACKGROUND

By 2023, approximately one in three households in the United States (U.S.) will be headed by individuals aged 65 and older (Joint Center for Housing Studies of Harvard University (Harvard Joint Center for Housing Studies, 2019). A large percentage of older adults prefer to remain in their homes and communities as they age (Clarke & Nieuwenhuijsen, 2009; Johnson & Appold, 2017), which promotes health by allowing people to live independently as long as possible and maintain connections with nearby family members, established social networks, community services, and/or other resources (Iecovich, 2014; Safran-Norton, 2010). Much of the literature in this area has focused on making homes more functional to address the increasing physical needs of aging populations (Alley et al., 2009; Safran-Norton, 2010; Szabo et al., 2017). In recent years, more attention has been placed on the financial aspects of housing and their influence on the health of Americans as they age (Engelhardt et al., 2013; Hamoudi & Dowd, 2014).

The combination of limited household budgets, declining household assets, and increasing cost of housing, has caused rising economic insecurity among seniors in recent years (Meschede et al., 2011). In 2015, approximately 4.8 million older U.S. households were spending over 50% of their incomes on housing costs (Molinsky & Airgood-Obrycki, 2018). Households that are housing insecure are at higher risk of foreclosure or eviction, and the resulting displacement from their homes and communities has negative health consequences (Harvard Joint Center for Housing Studies, 2019; Stahre et al., 2015). Pearson et al. (2019) estimate that by 2029, 54% of seniors will not have the financial resources needed to afford assisted living rent and medical out-of-pocket expenses. These projections were prior to the COVID-19 pandemic, where another economic recession may cause even more dire outcomes. Prior research has linked unaffordable housing to increased odds of poor self-rated health, hypertension, and arthritis (Pollack et al., 2010). Furthermore, unaffordable housing can negatively impact food access, and has been linked with non-adherence to healthcare and medication due to limited financial resources (Kirkpatrick & Tarasuk, 2011; Pollack et al., 2010; Stahre et al., 2015). These effects can be further complicated by tenure status. Compared to owners, renters tend to have fewer financial resources, making them vulnerable to increases in rent and other housing-related costs (Harvard Joint Center for Housing Studies, 2019). In 2016, homeowners over the age of 65 had an average net wealth of 319,200 USD compared to the net wealth of 6,700 USD for renters of the same age (Harvard Joint Center for Housing Studies, 2019). Homeownership provides opportunities to accumulate equity, placing homeowners at a greater advantage in terms of household wealth compared to renters. However, little is understood about how the negative health consequences of financial burden may differ between renters and homeowners.

It is projected that by 2035, 6.4 million renters and 11 million homeowners will be housing insecure (Harvard Joint Center for Housing Studies, 2019). While most older adults are homeowners, the proportion of renters has been increasing; from 2004 to 2020, the number of renters aged 55+ years had increased by 22%, comprising more than a quarter of existing renters (Harvard Joint Center for Housing Studies, 2020). Compared to older homeowners, older renters are more likely to report poor health and have higher levels of psychological distress and depression (Cairney & Boyle, 2004; Dalstra et al., 2006; Herbers & Mulder, 2017; Kavanagh et al., 2016). Older renters must also rely on housing providers or property managers to make accessibility renovations that allow them to age in place, while homeowners, if financially equipped, can better maintain their homes and make renovations on their own (Dietz & Haurin, 2003; Szabo et al., 2017). Therefore, it is no surprise that renters are more likely to become institutionalized for long-term care at a younger age than homeowners (Rouwendal & Thomese, 2013). More research on the potential effects of financial burden on the pathways linking tenure to health can help to focus attention and inform interventions that assist the aging population with housing and associated costs.

Although homeowners fare better when compared to renters, homeownership presents its own unique challenges, especially among older populations. Increasingly, homeowners are carrying mortgage debt into older age; between 1989 and 2018, the percentage of mortgage holders nearly doubled for those aged 65–74, and tripled for those 75+ years (Harvard Joint Center for Housing Studies, 2019). Even those without mortgage debt can still struggle to pay housing-related costs on retirement income, making them vulnerable to inflation and increased costs of healthcare and housing modification (Cairney & Boyle, 2004; Harvard Joint Center for Housing Studies, 2019). These circumstances highlight the need for more research that investigates how financial burden among homeowners with and without mortgage debt can affect the health of aging homeowners.

This study uses the Health and Retirement Study (HRS) to investigate associations of tenure status (renters and owners) and mortgage debt status, with financial burden and health among older adults in the U.S. This research will contribute to the literature in this area by generating evidence that is needed to inform programs and policies that can promote affordable and stable housing. To do this, our study sought to answer three questions: 1. Are tenure status and financial burden associated with health? 2. Are mortgage debt and financial burden associated with health? and 3. Are associations between tenure status, mortgage debt and health, moderated by financial strain or difficulty paying bills? For this study, health measures included self-rated health, heart condition diagnosis, and hospitalization in the last two years, which are all health conditions previously connected to housing in the general adult population (Burgard et al., 2012; Pollack et al., 2010; Sadowski et al., 2009). Based on the existing literature, we expect renters to have worse health outcomes than owners. We also expect that this relationship will be heightened by financial burden experienced. We expect similar findings among mortgage holders where they will fare worse in health compared to non-mortgage holders, with this relationship similarly worsened by financial burden.

METHODS

Data Sources

We used data from the HRS, a biennial survey of a cohort of noninstitutionalized U.S. adults 50 years and older. Details on the HRS sample design and validation of measurements have been previously published (Sonnega et al., 2014). We used data from the Leave-Behind Questionnaire (LBQ), which gathers more information on additional health outcomes, beyond that included in the basic survey. This study used cross-sectional data from participants who completed the LBQ during the 2012 and 2014 waves. We focus on these two years as they proceed the housing crash of 2008 with foreclosure rates highest in 2009 and 2010 (RealtyTrac, 2016). Since LBQ surveys were completed by half of the cohort in 2012, and the remaining half in 2014, data were aggregated resulting in a larger analytic sample size (n = 14,212). The sample was limited to respondents who identified as either renters or homeowners, had completed a previous wave (2010 or earlier, to account for baseline measures), had data available for health and financial variables and were 65 years and older, recognized by organizations like the Center for Disease Control and Prevention as “older persons” (Center for Disease and Control and Prevention, 2020). The total study sample included 6,449 unique respondents.

Dependent Variables

Three self-reported health outcomes were assessed: self-rated health (SRH), diagnosis of heart condition (HC), and hospitalization over the last two years. SRH is a common measurement to assess a person’s wellbeing and a reliable predictor for other health outcomes (Marshall & Tucker-Seeley, 2018). A five-point Likert scale measured SRH: 1 being excellent health and 5 being poor health. Hospitalization was dichotomized based on whether the respondent had a hospital stay in the past two years and HC was dichotomized based on whether the respondent was told by a doctor they had the health condition or not.

Independent Variables

Independent variables included tenure status, mortgage status, and financial burden measures. Respondents’ tenure was coded as either homeowners or renters, and included as a predictor variable for health among the total sample. Similarly, all homeowners were categorized as either having a mortgage (or any other loan that uses property as collateral), or not, used to assess mortgage debt as a health predictor among homeowners. Two measures of financial burden were used as independent variables: (1) Ongoing financial strain was dichotomized, based on participant’s response to being financially strained for 12 months or more or not; (2) Difficulty paying bills was dichotomized based on participant’s response to experiencing difficulties in paying monthly bills (“somewhat difficult,” “very difficult,” or “completely difficult.”) or not (“not at all” or “not very difficult”).These self-reported measures have been previously used to examine health outcomes among older populations including self-rated mental and physical health outcomes (Angel et al., 2003; Asebedo & Wilmarth, 2017)

Control Variables

All analyses were controlled for age, sex, race/ethnicity (non-Hispanic White (NHW), non-Hispanic Black (NHB), or Hispanic), education, cohabitation status, income-to-poverty ratio (IPR), ever smoked, high blood pressure (BP) diagnosis at baseline, and survey year. The IPR is an imputation from the RAND HRS data file, based on wealth and income compared to the household’s area income (Chien et al., 2013). To control for health behaviors and health status at baseline, smoking and high BP were used, respectively. Both measures are also underlying factors for the health outcomes in the study (Fuchs & Whelton, 2020; Singh & Yu, 2016).

Statistical Analyses

Statistical estimates were weighted, standard errors were adjusted for survey design, and were used in all analyses. We described socio-demographic characteristics for the full sample and compared these characteristics between tenure groups and mortgage status using t-tests (for means of two groups), chi-square tests (for frequencies of one or more categories) and their corresponding effect sizes using Cohen’s d and Cramer’s V, respectively, to examine differences across the groups. Survey ordinal regression analyses were used for models with SRH, while survey logistic regression analysis were used for HC and hospitalization. To answer our first research question examining whether tenure status and financial burden characteristics are associated with health, the total sample was used to regress each health outcome separately on tenure status (owner vs renter) and the financial burden variables. To answer our second research question examining whether mortgage debt and financial burden characteristics were associated with health, we regressed each of the health outcomes separately on mortgage debt and financial burden variables, respectively. For these models, the study sample included only homeowners. To answer our third research question, we used separate interaction terms to test for moderation effects between tenure status and each financial burden outcome, and for homeowners, between mortgage debt status and each financial burden outcome. Separate models were run for each of the health outcomes.

RESULTS

Table 1 presents the weighted descriptive characteristics of participants. Renters were on average older (76.2 years) compared to owners (73.8 years), while mortgage holders were younger (71.0 years) compared to non-mortgage holders (75.0 years). The sample consisted mostly of NHW participants (84.6%) followed by NHB (8.8%) and Hispanics (6.6%). While in the total sample and all homeowner subcategories a higher percentage lived with partners, among renters more (72.6%) were not living with a partner. Just over half of the sample were women, and with more female than male renters (63.4%). Approximately half of the total sample completed high school (54.2%), but a higher percentage of renters did not have a high school diploma (31.5%) compared to homeowners. The mean IPR was 5.1 for the whole sample, but lower (2.9) among renters compared to homeowners (5.5). Furthermore, the IPR was higher among mortgage holders (6.4) compared to non-mortgage holders (5.1). Financial strain was higher among renters (49.2%) compared to homeowners (31.9%), mortgage holders (46.9%), and non-mortgage holders (25.3%). Similarly, 36.6% of renters reported having difficulty paying bills, while less than a fifth of owners reported difficulties. Fifteen percent of non-mortgage holders reported difficulties paying bills while only 27.5% of mortgage holders reported similar difficulties. P-values of less or equal to 0.001 (0.01 for gender differences in mortgage and non-mortgage holders) indicated significant differences between owners and renters, and mortgage and non-mortgage holders for all the socio-demographic characteristics, while effect size ranged from 0.06 to 0.66 indicating small to medium strengths of association.

Table 1.

Characteristics of HRS participants by housing tenure and mortgage debt for variables reporting means.

	Total Sample (n = 6,449)	Renter (n = 966)	Owner (n = 5,483)	Mortgage Holders (n = 1,525)	Non-Mortgage Holders (n = 3,905)
Demographic Characteristics
Age (SE)	74.2 (0.11)	76.2 (0.35)	73.8 (0.12)	71.0 (0.16)	75.0 (0.14)
T value			6.41***		18.33***
Effect size			0.22		0.55
Race/Ethnicity (%)
Non-Hispanic White	84.6%	71.0%	86.0%	86.0%	87.5%
Non-Hispanic Black	8.8%	17.2%	8.2%	8.2%	6.9%
Hispanic	6.6%	11.8%	5.8%	5.8%	5.6%
X2 value			185.79***		16.81***
Effect size			0.17		0.06
Gender (%)
Male	44.5%	36.6%	46.0%	50.3%	44.5%
Female	55.5%	63.4%	54.0%	49.7%	55.5%
X2 value			28.98***		9.05**
Effect size			0.07		0.04
Cohabitation Status (%)
Lives with partner	61.5%	27.4%	67.7%	73.5%	65.1%
Does not live with partner	38.5%	72.6%	32.3%	26.5%	34.9%
X2 value			506.67***		24.50***
Effect size			0.28		0.07
Education (%)
No High School	19.3%	31.5%	17.0%	13.3%	18.6%
High School Completion	54.2%	51.2%	54.8%	50.5%	56.7%
Some college or above	26.5%	17.3%	28.2%	36.2%	24.7%
X2 value			103.62***		53.45***
Effect size			0.13		0.10
Mortgage Status (%)
Mortgage holder	NA	NA	30.5%	NA	NA
Non-mortgage holder	NA	NA	69.5%	NA	NA
Financial Characteristics
Income-to-poverty Ratio (SE)	5.1 (0.12)	2.9 (0.16)	5.5 (0.16)	6.4 (0.23)	5.1 (0.21)
T value			−8.63***		−4.29***
Effect size			−0.30		−0.13
Ongoing Financial Strain
Yes	34.8%	49.2%	31.9%	46.9%	25.3%
No	65.2%	50.8%	68.1%	53.1%	74.7%
X2 value			108.98***		234.20***
Effect size			−0.13		0.21
Difficulty Paying Monthly Payments
Difficult	21.6%	36.6%	18.9%	27.5%	15.0%
Not Difficult	78.4%	63.4%	81.1%	72.5%	85.0%
X2 value			148.89***		116.94***
Effect size			−0.15		0.15
Health outcomes
Self-rated health (5 = poor, 1 = best) (SE)	2.8 (0.02)	3.1 (0.04)	2.8 (0.02)	2.7 (0.03)	2.8 (0.02)
Heart condition	31.6%	36.8%	30.6%	28.2%	31.6%
Hospitalization	28.2%	34.6%	27.0%	26.6%	27.2%
High blood pressure at baseline	56.1%	61.8%	55.0%	54.6%	55.2%
Ever smoked	56.5%	62.2%	55.5%	59.7%	53.8%

^**p < .01

^***p < .001

For health outcomes, renters reported an average SRH score of 3.1 (5 being poor) compared with 2.8 among owners. The average percentages of those reporting having an HC diagnosis, and being hospitalized, for the entire sample were 31.6% and 28.2%, respectively. However, these proportions were higher among renters (36.8% and 34.6%). At baseline, 61.8% of renters reported having high BP, compared to 55.0% of owners; 54.6% of mortgage holders reported high BP, compared to 55.3% of non-mortgage holders. Over half of the sample (56.5%) reported having smoked at some point their lives, with 62.2% of renters having smoked compared to 55.5% of owners. Among owners, 59.7% of the mortgage holders reported ever smoking compared to 53.8% of non-mortgage holders.

Table 2 presents the results from the analyses addressing our first question, whether housing tenure, financial strain and difficulty paying bills are associated with health outcomes, accounting for covariates. The analyses were checked for multicollinearity and none was observed among the independent variables. Being a renter was associated with greater likelihood of poor SRH (OR = 1.26, p < 0.01) and hospitalization (OR = 1.26, p < 0.05) but not significantly associated with greater likelihood of HC diagnosis. Financial strain was significantly associated with a greater likelihood of poor SRH (OR = 1.43, p < 0.001), HC diagnosis (OR = 1.40, p < 0.001), and hospitalization (OR = 1.26, p < 0.010), while difficulty paying bills was significantly associated with greater likelihood of poor SRH (OR = 1.79, p < 0.001), and HC diagnosis (OR = 1.38, p < 0.010).

Table 2.

Health outcomes regressed on tenure status and financial burden outcomes

	Poor SRH		Heart Condition Diagnosis		Hospitalization (past 2 years)
	Odds Ratio	95% CI	Odds Ratio	95% CI	Odds Ratio	95% CI
Tenure
Owners (ref)	1.00		1.00		1.00
Renters	1.26**	1.07, 1.49	1.12	0.93, 1.35	1.26*	1.04, 1.51
Ongoing financial stress
No (ref)	1.00		1.00		1.00
Yes	1.43***	1.24, 1.65	1.40***	1.19, 1.65	1.26**	1.07, 1.49
Difficulty paying monthly payments
No (ref)	1.00		1.00		1.00
Yes	1.79***	1.51, 2.11	1.38**	1.14, 1.66	1.07	0.88, 1.30
Age
Age (continuous)	1.03***	1.02, 1.04	1.05***	1.04, 1.06	1.03***	1.02, 1.04
Race/Ethnicity
Non-Hispanic White (ref)	1.00		1.00		1.00
Non-Hispanic Black	1.44***	1.22, 1.71	0.82	0.65, 1.02	1.01	0.81, 1.25
Hispanic	1.94***	1.54, 2.46	0.67**	0.50, 0.91	0.82	0.62, 1.09
Gender
Male (ref)	1.00		1.00		1.00
Female	0.90	0.80, 1.01	0.67***	0.58, 0.76	0.91	0.80, 1.05
Cohabitation status
No (ref)	1.00		1.00		1.00
Yes	1.02	0.81, 1.01	1.08	0.93, 1.26	0.91	0.80, 1.05
Education
No high school (ref)	1.00		1.00		1.00
High school completion	0.54***	0.47, 0.63	1.14	0.96, 1.36	0.98	0.82, 1.16
Some college or above	0.35***	0.29, 0.42	0.90	0.73, 1.11	0.74**	0.60, 0.91
Income-to-poverty ratio
IPR (continuous)	0.99	0.98, 1.01	1.00	0.99, 1.01	1.00	0.99, 1.01
High blood pressure at baseline
No (ref)	1.00		1.00		1.00
Yes	1.76***	1.58, 1.96	1.74***	1.53, 1.99	1.54***	1.34, 1.76
Ever smoked
No (ref)	1.00		1.00		1.00
Yes	1.26***	1.13, 1.40	1.25**	1.09, 1.42	1.18*	1.03, 1.34
Cohort year
2012 (ref)	1.00		1.00		1.00
2014	1.11	0.99, 1.24	0.98	0.86, 1.11	0.86*	0.76, 0.98

^*p < .05

^**p < .01

^***p < .001. All models were controlled for age, race, education, IPR, cohabitation status, baseline high BP, ever smoked and survey year.

Results from analyses assessing our second research question, the extent to which mortgage debt status, financial strain, or difficulty paying bills are associated with each health outcome, are shown in Table 3. Controlling for covariates, having mortgage debt was associated with lower likelihood of poor SRH (OR = 0.80, p < 0.010). Regardless of mortgage status, financial strain was associated with greater likelihood of poor SRH (OR = 1.57, p < 0.001), HC diagnosis (OR = 1.37, p < 0.010) and hospitalization (OR = 1.29, p < 0.010); difficulty paying bills was associated with greater likelihood of poor SRH (OR = 1.83, p < 0.001), and HC diagnosis (OR = 1.50, p < 0.001).

Table 3.

Health outcomes regressed on mortgage status and financial burden outcomes.

	Poor SRH		Heart Condition Diagnosis		Hospitalization (past 2 years)
	Odds Ratio	95% CI	Odds Ratio	95% CI	Odds Ratio	95% CI
Mortgage status & financial characteristics
Mortgage status
Non-mortgage holder (ref)	1.00		1.00		1.00
Mortgage holder
Ongoing financial stress	0.80**	0.70, 0.92	0.89	0.76, 1.05	1.06	0.89, 1.25
No (ref)	1.00		1.00		1.00
Yes	1.57***	1.34, 1.83	1.37**	1.14, 1.63	1.29**	1.07, 1.55
Difficulty paying monthly payments
No (ref)	1.00		1.00		1.00
Yes	1.83***	1.52, 2.21	1.50***	1.21, 1.85	1.02	0.81, 1.27
Demographic characteristics
Age
Age (continuous)	1.03***	1.02, 1.04	1.05***	1.04, 1.06	1.03***	1.02, 1.04
Race/Ethnicity
Non-Hispanic White (ref)	1.00		1.00		1.00
Non-Hispanic Black	1.46***	1.20, 1.77	0.81	0.63, 1.05	1.04	0.81, 1.32
Hispanic	1.71***	1.30, 2.24	0.74	0.53, 1.05	0.77	0.55, 1.09
Gender
Male (ref)	1.00		1.00		1.00
Female	0.89	0.79, 1.01	0.66***	0.57, 0.77	0.86*	0.74, 0.99
Cohabitation status
No (ref)	1.00		1.00		1.00
Yes	1.09	0.95, 1.24	1.06	0.90, 1.25	1.08	0.91, 1.27
Education
No high school (ref)	1.00		1.00		1.00
High school completion	0.58***	0.49, 0.68	1.18	0.97, 1.44	1.05	0.87, 1.28
Some college or above	0.38***	0.31, 0.46	0.93	0.74, 1.17	0.77	0.61, 0.97
Income-to-poverty ratio
IPR (continuous)	0.99	0.98, 1.01	1.00	0.99, 1.00	1.00	0.99, 1.01
Medical history and wave characteristics
High blood pressure at baseline
No (ref)	1.00		1.00		1.00
Yes	1.82***	1.62, 2.05	1.83***	1.59, 2.12	1.56***	1.33, 1.79
Ever smoked
No (ref)	1.00		1.00		1.00
Yes	1.26***	1.12, 1.42	1.20*	1.04, 1.38	1.10	0.95, 1.27
Cohort year
2012 (ref)	1.00		1.00		1.00
2014	1.13*	1.01, 1.27	1.03	0.90, 1.18	0.86	0.75, 1.00

^*p < .05

^**p < .01

^***p < .001. All models were controlled for age, race, education, IPR, cohabitation status, baseline high BP, ever smoked and survey year.

Table 4 shows results from models assessing our third research question, whether the associations between tenure status or mortgage status and health were moderated by the financial burden variables. Results are not consistent with the hypothesis that the relationships between tenure status or mortgage status, and the health outcomes are dependent on either financial strain or difficulty paying monthly bills. They suggest that tenure status and mortgage status remain predictors of poor SRH and hospitalization, and that these relationships are consistent regardless of financial burden experienced.

Table 4.

Modification of the association between tenure status and health outcomes by: financial stress (Model 1a) and difficulty paying bills (Model 1b); and modification of the association between mortgage debt and health outcomes by: financial stress (Model 2a) and difficulty paying bills (Model 2b) Health outcomes regressed on tenure status (model 1), mortgage status (model 2) and financial burden outcomes.

	Poor SRH		Heart Condition Diagnosis		Hospitalization (past 2 years)
	Odds Ratio	95% CI	Odds Ratio	95% CI	Odds Ratio	95% CI
Model 1a
Renter	1.37**	1.08, 1.74	1.13	0.88, 1.44	1.25	0.97, 1.60
Has ongoing financial strain	1.47***	1.27, 1.70	1.40***	1.18, 1.66	1.26*	1.05, 1.50
Renter*Has financial strain	0.84	0.61, 1.16	1.00	0.70, 1.42	1.02	0.71, 1.46
Model 1b
Renter	1.27*	1.03, 1.55	1.24	0.99, 1.55	1.19	0.95, 1.48
Has difficulty paying monthly payments	1.79***	1.49, 2.14	1.47***	1.20, 1.80	1.03	0.83, 1.27
Renter*Has difficulty paying	0.99	0.71, 1.39	0.75	0.52, 1.10	1.18	0.81, 1.73
Model 2a
Mortgage holder	0.79**	0.66, 0.94	0.94	0.67, 1.05	1.05	0.84, 1.31
Has ongoing financial strain	1.54***	1.30, 1.84	1.30**	1.06, 1.60	1.28*	1.03, 1.58
Mortgage holder*Has financial strain	1.04	0.79, 1.37	1.15	0.83, 1.59	1.02	0.73, 1.41
Model 2b
Mortgage holder	0.82**	0.70, 0.95	0.96	0.80, 1.16	1.04	0.86, 1.27
Has difficulty paying monthly payments	1.88***	1.50, 2.36	1.68***	1.31, 2.15	1.00	0.77, 1.30
Mortgage holder*Has difficulty paying	0.94	0.68, 1.29	0.75	0.52, 1.08	1.05	0.72, 1.52

^*p < .05

^**p < .01

^***p < .001. All models were controlled for age, race, education, IPR, cohabitation status, baseline high BP, ever smoked and survey year.

DISCUSSION

Our findings with respect to our first question are consistent with findings by others, suggesting that renters have poorer health than owners in both the senior (Johnson & Appold, 2017) and the general populations (Pollack et al., 2010). Renting is linked to less financial security and economic opportunity, housing instability and inability to accumulate wealth through equity building, among middle-aged adults and is likely working through similar mechanisms in older populations (Rohe et al., 2002; Shlay, 2006). Furthermore, the residential stability homeownership provides could translate to stronger social networks within the neighborhood resulting in better health generally (Rohe & Stewart, 1996). Our findings also indicate that regardless of financial burden, housing tenure is a predictor of self-rated health and hospitalization, consistent with our hypothesis that housing is a social determinant of health.

Regardless of tenure status, adults with a financial burden (as measured by ongoing financial strain and difficulty paying bills) were more likely to experience poor self-rated health and heart condition diagnoses. Furthermore, experiences of ongoing financial strain were also associated with hospitalization. Continuing or rising housing costs including mortgage payments, rents and property taxes, along with age-related declines in household income, and increasing health costs among older adults may contribute to financial stress and in turn harm health. This finding is consistent with the prior literature which indicates that unaffordable housing costs and the risk or reality of displacement through foreclosures or evictions can harm physical and mental well-being (Cagney et al., 2014; Pollack et al., 2010; Pollack & Lynch, 2009; Safran-Norton, 2010). Such effects may be especially problematic in strong housing markets where affordable housing does not meet growing demands, making it more difficult to keep up with housing-related costs (Joint Center for Housing Studies of Harvard University, 2019). Research considering the multi-level impacts of individual economic and health outcomes together with neighborhood or city-level housing market predictors could provide further insight on areas of growing concern in terms of affordability for seniors and others with limited income resources.

Having a mortgage was only associated with being more likely to report better self-rated health. While we controlled for age, it is possible that this finding may reflect that the average age of mortgage holders was 71, an average of four years younger than those without a mortgage and about three years younger than the total sample. Compared to non-mortgage holders, mortgage holders reported on average slightly better self-rated health and lower proportions of heart condition diagnosis, hospitalization rates and high blood pressure at baseline. However, regardless of mortgage status, ongoing financial strain and difficulties paying bills were associated with poor self-rated health and health condition diagnosis, while hospitalization was associated with ongoing financial stress. Similar to the tenure analysis, these findings are consistent with the large body of the existing literature showing the negative health impacts of a financial burden on aging populations, and highlighting the potential for health benefits with increased financial resources and opportunities for this population in the U.S. (Huguet et al., 2008; Krause et al., 2008; Litwin & Sapir, 2009; Schoeni et al., 2005).

Financial burden measures did not modify the relation between mortgage debt and health. Although not having a mortgage could reduce housing expenses considerably, other costs like property taxes, utilities, and home insurance could still be a burden for those facing a financial burden and are more likely to live on fixed incomes. These strains could also translate to limited or no resources for making needed home repairs (e.g., replacement of furnace or roof repair), which in turn could impact health and lower one’s quality of life. Although the HRS does cover questions on home modifications to accommodate wheelchairs, questions on home repairs needed (e.g., furnace replacement or roof repairs) or repairs made are not covered, making it difficult to capture these housing-related costs. Future research could benefit from considering questions on housing satisfaction or money spent on repairs, to approximate for these associated costs.

Finally, our results could have captured the lingering effects on housing values post-Great Recession from 2007 to 2009. These effects could result in homeowners being “locked in place” or not able to move since their home values did not recover, especially for those in distressed neighborhoods (Gustman et al. 2012; Engelhardt et al., 2013). Such connections highlight the need for research that considers the multi-level health effects of neighborhood- and individual-level housing characteristics to examine in what circumstances aging in place may have negative health consequences.

The HRS is one of the few nationally representative samples of older populations that allows for individual-level analyses of health, housing status, and financial burden data. Still, a weakness of this study is the cross-sectional design that makes it difficult to assess causal relationships between housing tenure and health. Adjustments were made for poverty to mitigate the concerns that people living in poverty may have worse health. Furthermore, adjustments were made for baseline blood pressure and ever having smoked to account for previous health status. While the HRS has longitudinal data, it focuses on those aged 50 and over, limiting the opportunity to examine health and housing status over the life course that could significantly impact the ability to purchase a home or pay off mortgage debt in later life. Interventions, like the ones presented below, are designed to address the problems we have highlighted, and would be applicable despite the direction of causality. Furthermore, to explore these relationships further, we plan to use future waves of HRS data to establish a temporal ordering of events to address some of these questions. We also did not examine how neighborhood context (e.g., housing prices, physical conditions) may interact with tenure status and financial burden to influence health. In the third paper of this series, we will combine neighborhood-level data to study whether certain neighborhood factors could mitigate or exacerbate the effects of financial burden for renters and homeowners. Furthermore, residual confounding is always a potential source of bias in studies of this nature. Our study considered only self-reported measures of financial burden. Still, these variables provide insight on perceived stress, which has been linked to all the outcomes considered in the study. Finally, the HRS sample is predominantly NHW, and we limited our sample to only NHWs, NHBs, and Hispanics. Therefore, findings may not be generalizable across all racial/ethnic groups in the U.S. Given historic patterns of housing discrimination and the persistence of racially stratified dual housing markets in the U.S., future research may consider whether factors like housing tenure and financial strain may contribute to long-standing racial/ethnic disparities in health.

Despite these limitations, our findings suggest that financial burden may negatively influence the health of older adults, regardless of household tenure and mortgage debt, raising important implications for future programs and policies targeted at older populations. Few U.S. programs exist to assist with housing costs among older populations and of those that do have requirements that exclude certain populations. For example, Medicaid’s Money Follows the Person program provides up to 45,000 USD for housing modifications for nursing home residents returning to their homes, but is restricted to Medicaid recipients which limits participation based on income and assets (Medicaid.gov n.d.). Other programs like the U.S. Department of Energy’s Weatherization Assistance Program assist low-income homeowners of all ages with the associated costs for weather resistance modifications. As constructed, these programs benefit homeowners but are less effective in reaching renters living in homes with poor energy qualities (U.S. Department of Energy, n.d.). Through private entities, programs like reverse mortgage or home equity line of credit, can help pay for housing modifications, these loans have limited flexibility for uncertain life events that could arise resulting in large debts to the borrowers (Kutty, 1998). Finally, programs aimed at alleviating costs like property taxes for older fixed-income homeowners can come with great barriers in submitting applications and required documents (Eisenberg et al., 2019). Furthermore, these findings suggest the need for more opportunities to purchase homes earlier in life to prevent housing instability later on. Findings from our research and others examining financial burden and housing status can be leveraged to inform policies and programs that protect older populations – regardless of whether they own or rent – to promote housing affordability and increase opportunities for older adults to age healthfully in their homes and communities.

CONCLUSION

This study explored the relationship between tenure status, mortgage status, financial burden, and health. Among homeowners, we were able to examine and compare the effects of financial burden on health within mortgage debt holders and non-mortgage debt holders, two subgroups often overlooked in the analyses. Findings contribute important new insight on housing tenure and mortgage debt as significant determinants of health. Our results highlight the need for more robust and inclusive policies and programs aimed at assisting older populations remaining in their homes and communities.

Biography

Dr. Roshanak Mehdipanah PhD is an Assistant Professor in the Department of Health Behavior and Health Education in the School of Public Health. Dr. Mehdipanah has led several projects on housing and health including evaluations of existing housing policies and their impacts on health inequities. She has also examined associations between neighborhood characteristics and housing discrimination in the Detroit Metropolitan Area. She specializes in innovative research methods including realist evaluations and concept mapping to develop conceptual frameworks linking complex interventions to health. Dr. Mehdipanah's current research portfolio focuses on various aspects of urban health including urban renewal, design, housing, and gentrification.

Jaclyn Martin MPH is a project manager on the Advancing Health Equity: Leading Care, Payment, and Systems Transformation initiative housed at the University of Chicago and funded by the Robert Wood Johnson Foundation. She holds a master's degree in public health from the Department of Health Behavior and Health Education in the School of Public Health at the University of Michigan. Her education focused on the impact of social policies on health inequities. She worked on this project with Dr. Mehdipanah and the rest of the team while completing her MPH.

Alexa K. Eisenberg MPH is a doctoral candidate in the Department of Health Behavior and Health Education at the School of Public Health and a National Institute of Health trainee at the Population Studies Center at the University of Michigan. Their researchintegrates methods from demography, social geography, and legal epidemiology to understand how housing and urban policies reproduce racialized hierarchies in population health in US metropolitan areas. Alexa is involved with several evaluation studies that focus on property tax foreclosure and eviction prevention in Detroit, Michigan.

Dr.Amy J. Schulz PhD is University Diversity and Social Transformation Professor in the Department of Health Behavior and Health Education at the University of Michigan School of Public Health. Her research focuses on social factors that contribute to health with a particular focus on social and physical environmental conditions and their effects on health and health equity with a focus on urban health. A majority of Dr. Schulz's research is conducted with partners in Detroit, using a community-based participatory research approach. She has been involved in working with Detroit partners to understand and address factors that contribute to excess risk of cardiovascular disease in Detroit, conduct health impact assessments of proposed policies, and develop public health action plans to reduce air pollution and promote health in Detroit and the surrounding area.

Dr. Lewis B. Morgenstern MDis professor of neurology, emergency medicine and neurosurgery at the University of Michigan Medical School. He also is professor of epidemiology in the Center for Social Epidemiology and Population Health at the UM School of Public Health. Dr. Morgenstern is a National Institutes of Health-funded principal investigator of studies that aim to reduce stroke and cognitive health disparities with respect to race, ethnicity, and gender. His other research focus is the treatment of intracerebral hemorrhage and mobilizing health care professionals and communities to treat acute ischemic stroke. He is currently principal investigator of five NIH R01 grants.

Dr. Kenneth M. Langa MD PhD is the Cyrus Sturgis Professor in the Department of Internal Medicine and Institute for Social Research, a Research Scientist in the Veterans Affairs Center for Clinical Management Research, and an Associate Director of the Institute of Gerontology, all at the University of Michigan. He is also Associate Director of the Health and Retirement Study (HRS), a National Institute on Aging funded longitudinal study of 20,000 adults in the United States. He is a General Internist with an active clinical practice treating adult patients, an elected member of the American Society for Clinical Investigation (ASCI), and an elected fellow of the American Association for the Advancement of Science (AAAS).