This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
Author's personal copy
The high price of debt: Household financial debt and its impact
on mental and physical health
Elizabeth Sweeta,*, Arijit Nandib, Emma K. Adamc,d, Thomas W. McDaded,e
aNorthwestern University Feinberg School of Medicine, Department of Medical Social Sciences, Abbott Hall, 710 N. Lake Shore Drive, Suite 729,
Chicago, IL 60611, USA
bMcGill University, Institute for Health and Social Policy, Meredith Charles House, 1130 Pine Ave West, Montreal, Quebec H3A 1A3, Canada
cNorthwestern University, School of Education and Social Policy, Annenberg Hall Room 108, 2120 Campus Drive, Evanston, IL 60208, USA
dNorthwestern University, Cells to Society (C2S): The Center on Social Disparities and Health, at the Institute for Policy Research, 2040 Sheridan Road,
Evanston, IL 60208, USA
eNorthwestern University, Department of Anthropology, 1810 Hinman Ave, Evanston, IL 60208, USA
a r t i c l e i n f o
Available online 16 May 2013
Social determinants of health
a b s t r a c t
Household financial debt in America has risen dramatically in recent years. While there is evidence that
debt is associated with adverse psychological health, its relationship with other health outcomes is
relatively unknown. We investigate the associations of multiple indices of financial debt with psycho-
logical and general health outcomes among 8400 young adult respondents from the National Longitu-
dinal Study of Adolescent Health (Add Health). Our findings show that reporting high financial debt
relative to available assets is associated with higher perceived stress and depression, worse self-reported
general health, and higher diastolic blood pressure. These associations remain significant when con-
trolling for prior socioeconomic status, psychological and physical health, and other demographic factors.
The results suggest that debt is an important socioeconomic determinant of health that should be
explored further in social epidemiology research.
? 2013 Elsevier Ltd. All rights reserved.
It is difficult to miss the growing impact of financial debt in the
everyday lives of Americans. Since the 1980s overall debt in
American households has tripled (Harvey, 2010). Between 1989 and
2006, total consumer credit card debt rose from $211 billion to$876
billion (2006 dollars), and the proportion of indebted households
carryingover $10,000 in creditcard debt rose from3% to27% (Garci,
2007). Home foreclosureshave also skyrocketed; recent foreclosure
rates are nearly 5 times higher than at any other time since 1979
(Gruenstein Brocian, Wei & Ernst, 2010). And, widely publicized in
March of 2012, Americans’ student loan debt recently surpassed $1
trillion (Mitchell & Jackson-Randall, 2012). Concomitant with these
rises in debt, credit industry deregulation, including key legislative
decisions in 1978 and 1996, have led to widespread increases in
loan fees and interest rates and a relaxation of loan granting con-
ditions (Garcia, 2007). The resulting “democratization” of credit
availability has meant that segments of the American population
previously excluded from many forms of credit now have more
equal opportunities to accumulate debt. In short, financial debt has
become a fixture of the American household economy.
Despite the growing significance of indebtedness in the eco-
nomic lives of Americans, financial debt is largely neglected in
research on social and economic determinants of health. This is
surprising since debt is clearly an increasingly important category
of socioeconomic experience. The link between socioeconomic
status (SES) and health has long been recognized (Adler et al.,1994;
Adler & Stewart, 2010), but sotoohave the inadequacies of standard
SES indices e income, education, and occupation e to fully capture
the meaning of socioeconomic experience (Adler, 2009; Braveman
et al., 2005; Shavers, 2007; Sweet, 2011). Other factors are impor-
tant constituents of socioeconomic position, such as wealth, assets,
symbolic capital and, notably, debt, but these are rarely considered
in health research(Adler,2009; Sweet, 2011). Drentea and Reynolds
(2012) recently made a call for greater attention to household debt
in work on the social determinants of health. Here we echo that
sentiment and offer evidence that debt is indeed an important
predictor of health outcomes.
To date,existing research on thehealthconsequencesof debthas
focused largely onpsychological health. Historical and ethnographic
* Corresponding author. Tel.: þ1 312 503 0352.
E-mail addresses: email@example.com (E. Sweet), firstname.lastname@example.org
(A. Nandi), email@example.com (E.K. Adam), firstname.lastname@example.org
Contents lists available at SciVerse ScienceDirect
Social Science & Medicine
journal homepage: www.elsevier.com/locate/socscimed
0277-9536/$ e see front matter ? 2013 Elsevier Ltd. All rights reserved.
Social Science & Medicine 91 (2013) 94e100
Author's personal copy
Dudley, 2000; Graeber, 2011; Williams, 2005, 2008). And indeed
several empirical studies have found that financial strains such as
personal debt and home foreclosures are strong predictors of
depression, general psychological distress, mental disorders, and
suicidal ideation and behavior (Bridges & Disney, 2010; Brown,
Taylor & Price, 2005; Drentea & Reynolds, 2012; Jenkins et al.,
2008; McLaughlin et al., 2011; Meltzer et al., 2011; Pollack & Lynch,
2009; Reading & Reynolds, 2001; Selenko & Batinic, 2011). These
findings highlight the psychological potency of being indebted and,
as others have noted, have implications for other health conse-
quences of debt (Dossey, 2007; Drentea & Lavrakas, 2000).
Psychosocial factors, including stress and its mental health
correlates like depression and anxiety, are thought to be key
mechanisms through which SES “gets under the skin” to impact
health and health disparities, and a substantial body of work has
now investigated this pathway (reviewed in: Matthews & Gallo
(2011)). The experience of stress is known to lead to short- and
long-term physiological changes that play key roles in several
disease processes, particularly those involving metabolic and car-
diovascular systems (McEwen, 2004). Stress can also impact health
indirectly by influencing health behaviors, including diet, physical
activity, and substance use (McEwen & Seeman, 1999). Therefore,
considering the ample evidence that debt has psychological con-
sequences, it is possible that it could also impact other health
outcomes through psychosocial pathways.
Remarkably few studies, however, have explicitly examined
financial debt in relation to health. Drentea and Lavrakas (2000), in
an Ohio-based study of credit card debt, stress, and health, found
debt-to-income ratio and debt-related stress to be associated with
worse self-reported health and physical functioning. Subsequent
studies have explored debt only indirectly, finding that debt-related
financial stress is associated with worse self-reported health (Kim
et al., 2003; O’Neill et al., 2006), and that clients of credit coun-
seling services have higher odds of being overweight or obese than
the general population (Munster, Ruger, Ochsmann, Letzel &
Toschke, 2009). Clearly, more work is needed to confirm and better
understand the relation of financial debt to health, including clari-
investigating their impact on a broader array of health outcomes.
Moreover there is a need to more fully account for prior condi-
tions and experiences that may complicate the association of debt
with health. Drentea and Lavrakas (2000) have suggested that debt
“may be a more sensitive barometer of financial well-being than
income” because it represents accumulated hardships over time.
While this observation reinforces the likely importance of debt as a
socioeconomic indicator, it also points to the potential confounding
that arises from longitudinal accumulation of debt. Personal finan-
cial debt may result from acute life events, such as job loss, divorce,
or medical emergencies, which may themselves be psychosocial
stressors orhealth determinants. Accounting for prior psychosocial,
socioeconomic and health conditions is therefore critical for un-
derstanding the relationship between financial debt and health.
In this paper we aim to bring greater attention to debt as a social
determinant of health by exploring its association with both psy-
chological and general health outcomes. Specifically, we explore
the relationship of personal financial debt with health outcomes in
young adults (aged24e32 years) in the National Longitudinal Study
of Adolescent Health (Add Health). This nationally representative
cohort study has followed participants for over 15 years and is thus
an excellent data source for examining the association of debt with
health while accounting for prior conditions and events. Further,
the Add Health study contains data on multiple indices of debt,
including subject-reported absolute levels of household debt as
well as perceptions of debts relative to assets. We therefore explore
the association of multiple measures of debt with health outcomes
in order to isolate dimensions of debt that matter most for health.
Furthermore, we explore the association of debt with multiple
measures of health: to establish consistency with prior findings, we
examine two outcomes related to psychological health e perceived
stress and depressive symptoms; we also expand our investigation
to test the impact of debt on three additional health outcomes e
self-reported general health, systolic and diastolic blood pressure.
By exploring basic health impacts of financial debt, we aim to
establish a baseline from which future research can further inves-
tigate this relatively neglected socioeconomic health determinant.
Study design and sample
Add Health is a nationally representative cohort study with four
waves of in-home interviews conducted since its initiation in
school year 1994/1995. At Wave I, approximately 20,000 adoles-
cents in grades 7e12 participated in in-home interviews. A clus-
tered sampling design was utilized, in which an 80-school sample
was selected that was nationally representative in terms of
ethnicity, urbanicity, school size, type, and US region. Within these
schools, students were randomly chosen within grade and sex
strata. Some sub-groups of students were over-sampled, including
African American adolescents from higher-educated families. Wave
II was conducted one year after Wave I. While 15,000 of the original
respondents were re-interviewed at Wave II, those who had grad-
uated were not included. Wave III data collection took place in
2002/2003 and included 15,170 of the original respondents plus
1507 partners of those respondents. The most recent wave of data
collection (Wave IV) took place in 2007/2008 when the cohort was
24e32 years old, and included 15,701 of the original respondents.
In all waves in-home interview datawas collected via computer-
(CAPI/CASI). Interview questions covered participant demographic
and socioeconomic conditions, as well as psychological and general
health, health services use, behavior, and extensive social rela-
tionship information. In Wave IV measures of blood pressure and
other cardiovascular and metabolic biomarkers were introduced, as
were questions about household and personal financial debt.
In this paper we restrict our analyses to data from Waves I, III,
and IV. All key independent (debt) and dependent variables (health
outcomes) are from Wave IV. Data on sociodemographic, psycho-
social, and general health from Waves I and III which could influ-
ence both debt and health status at Wave IV are also included in
analyses as control variables. Since Wave II was conducted only one
yearafter Wave I and excluded all graduating seniors, wedo notuse
data from that wave. All original Wave I respondents were eligible
to participate at Wave IV, and therefore the most recent wave of
data collection includes some respondents who did not participate
in Wave III. Our analytic sample is thus smaller than the full Wave
IV sample since it includes only those respondents who partici-
pated in each of the three waves of data that we are using (Waves I,
III, and IV) and who are not missing sampling weight information
(n ¼ 9421). Further, we exclude respondents with missing data on
key independent and dependent variables (described below). Our
final analytic sample includes 8400 respondents. Approval was
obtained from the Northwestern University Institutional Review
Board to conduct secondary analyses of the Add Health data.
Personal financial debt was measured at Wave IV in two ways.
First, a subjective assessment of net status asked respondents to
E. Sweet et al. / Social Science & Medicine 91 (2013) 94e100
Author's personal copy
“suppose you and others in your household were to sell all of your
major possessions (including your home), turn all of your in-
vestments and other assets into cash, and pay off all of your debts.
Would you have something left over, break even, or be in debt?” We
dichotomized this variable to represent those who said that they
would still be in debt after liquidating all assets (i.e. had high sub-
jective debt-to-asset ratio, or relative debt) versus those who would
not still be in debt. Second, respondents were asked to “Think about
and others in your household owe altogether? Include all debts,
including all types of loans, credit card debt, medical or legal bills,
or more”. To ease interpretation responses were transformed to
reflect the midpoint of each category and rescaled to tens of thou-
sands of dollars. In addition to these two measures, total household
assets (excluding home equity) were reported in dollars. The ratio of
as an additional independent variable in analyses.
Perceived stress was measured at Wave IV with four items from
the Cohen Perceived Stress Scale (Cohen, Kamarck & Mermelstein,
1983). The questions measure respondents’ feelings of stress and
lack of control over the preceding month. Summed scores across
the four questions range from 0 to 16.
Symptoms of depression were measured at all waves using
questions from the Center for Epidemiologic Studies Depression
(CES-D) Scale (Radloff, 1977). The number of CES-D questions
administered at each wave of data collection varied; at Wave I a 19-
item version was used, at Wave III a 9-item version, and at Wave IV
a 5-item version was administered. All questions assess symptoms
of depression experienced during the preceding week. For each
wave responses were reverse coded as appropriate and summed to
create a total depressive symptoms score.
question in which participants rated their health on a five-point
scale. Response options ranged from “excellent” (1) to “poor” (5).
Higher scores on this scale thus indicate worse general health.
Systolic and diastolic blood pressure was measured at Wave IV
using a Microlife BP3MC1-PC-IB oscillometric blood pressure
monitor. Respondent arm circumference was measured prior to
blood pressure readings to ensure that appropriate cuff sizes were
used with the monitor. Respondents rested in a seated position for
5 min, after which field interviewers administered three consecu-
tive systolic and diastolic readings at 30-s intervals. The average of
the second and third of these readings was used to construct sys-
tolic and diastolic blood pressure variables (Entzel et al., 2009).
Wave I psychological, health, and socio-demographic variables
Participants reported socio-demographic information as well as
general and psychological health problems and risk factors at
Wave I (during adolescence) that could influence both health and
debt status later in life (at Wave IV). Health conditions and risk
factors measured at Wave I include: regular smoking (at least one
cigarette per day for 30 days), presence of major physical limita-
tions (difficulty using hands, arms, legs, or feet because of a per-
manent physical condition), physical activity (a summary measure
of how often they reported engaging in various activities, such as
bicycling, dancing, and playing sports, in the previous week), self-
reported general health (described above), depressive symptoms
(measured by the CES-D, described above), and dietary quality
(how often they ate fruits, vegetables, and sweets during the day
prior to the Wave I interview). Participants also reported their race
(white, black, Asian, Native American, or Other) and Hispanic
ethnicity, as well as whether or not they had a physical medical
exam in the preceding year. Self-reported height and weight were
used to calculate participant body mass index (BMI, kg/m2) at Wave
I. In addition to information reported by participants themselves,
parents of participants reported their highest educational attain-
ment (coded as less than high school, high school, some college,
college degree, and post-college).
Wave III psychological, health, and socio-demographic variables
At Wave III participants again reported health and demographic
information that could be associated with future health and eco-
nomic status, including: number of hospitalizations in the pre-
ceding 5 years, whether they smoke regularly, depressive
symptoms, general health status, number of diseases ever diag-
nosed with, exercise in the previous week, whether they had a
physical medical exam in the past year, the number of months in
the past year they had health insurance, and whether they had
skippedmedical carewhen theyneeded it forany reasonin the past
year. Self-reported height and weight were again used to calculate
BMI. Participants also reported their highest level of education,
their household income (in dollars), and whether they owned or
were buying a home.
Wave IV socio-demographic variables
Wave IV annual household income was reported in dollars (less
than $5000 through $150,000 or more) and re-scaled to reflect
thousands of dollars. Other self-reported variables include highest
level of education attained, age in years, sex, the number of adults
and children living in the household, whether they had ever been
married, whether they smoke regularly, their level of regular
physical activity, the number of months in the past year they had
health insurance, whether they had lost a job in the previous 5
years, and whether they own their home.
Descriptive statistics, including means, standard errors, ranges
and percentages were calculated for key independent (Wave IV
subjective relative debt, absolute debt, and calculated debt-to-asset
ratio) and dependent (Wave IV perceived stress, depressive
symptoms, general health, SBP and DBP) variables as well as Wave
IV socio-demographic factors (age, race/ethnicity, sex, household
income, assets, education, health insurance, and home ownership).
Relationships among independent variables were assessed using
Pearson’s coefficients of pairwise correlations.
For each dependent variable, a series of ordinary least squares
(OLS) multiple regression models were run. Model 1 tests the un-
adjusted association of each debt variable with each dependent
variable. In Model 2 all Wave I and Wave III socio-economic, psy-
chological, and health factors (general health, depressive symp-
toms, medical exams, health insurance, skipped medical care,
disease diagnoses, hospitalizations, exercise, smoking, diet, BMI,
income, education, parental education, home ownership, and race/
E. Sweet et al. / Social Science & Medicine 91 (2013) 94e100
Author's personal copy
ethnicity) that could influence the association between debt and
health at Wave IV are added. In Model 3 all concurrent Wave IV
socioeconomic and demographic factors (number of people in the
household, education, income, smoking, physical activity, marital
status, health insurance, job loss, and home ownership) that could
confound the association of debt with health are added (for
calculated debt-to-asset ratio models, the main effects of Wave IV
household debts and assets are omitted). All descriptive statistics
and analyses were conducted using STATA 11.0 (StataCorp., 2009)
and weighted using appropriate sampling weights to account for
the complex sample design, selection, and non-response.
Table 1 presents descriptive statistics for the key Wave IV pre-
dictors, health outcomes, and socio-demographic variables. This
young adult sample (average age just under 28 years) was 47% male
and 70% non-Hispanic white. Respondents’ mean household in-
come was reported at just over $62,000 and mean assets were
$85,500. Mean non-mortgage household debts were just under
$30,000, and the mean calculated ratio of household debts to assets
was 2.5. Just over 20% of respondents reported that they would still
be in debt if they liquidated all of their assets (i.e. have high sub-
jective relative debt).
Correlations among independent variables and other relevant
Wave IV socioeconomic indicators are summarized in Table 2. All
debt-related and socioeconomic variables are significantly corre-
lated, with the strongest correlation between household income
and assets (r ¼ 0.423). Of note, absolute household debt is posi-
tively correlated with other socioeconomic indicators, such that
those with higher household incomes and assets also have higher
household debt. Also, subjective debt-to-asset ratio and calculated
debt-to-asset ratio are significantly but not highly correlated
(r ¼ 0.338).
Table 3 presents results from OLS regressions of total household
debts, calculated debt-to-asset ratio, and subjective debt-to-asset
ratio on health outcomes. Total household debt is an inconsistent
predictor of outcomes across the three models. In unadjusted
models (Model 1), higher total debt is not significantly associated
with any outcome, except for better self-reported general health.
With the addition of Wave I and Wave III control variables, asso-
ciations between household debt and several health outcomes
change direction, and with the addition of other Wave IV
socioeconomic variables total household debt becomes a significant
predictor of higher perceived stress and depressive symptoms, and
worse self-reported general health.
In unadjusted models (Model 1), a higher calculated ratio of
household debts compared to assets is associated with higher
perceived stress and depressive symptoms and worse general
health, but is not associated with either systolic or diastolic blood
pressure. With the addition of Wave I, Wave III, and Wave IV control
variables (Models 2 and 3), associations with perceived stress,
depression and general health remain significant.
Reporting high subjective debt-to-asset ratio (still being in debt
after liquidating all assets) is associated with significantly higher
perceived stress and depressive symptoms and worse self-reported
general health in unadjusted models (Model 1). Additionally, while
not associated with SBP, high subjective debt-to-asset ratio is
associated with significantly higher DBP in unadjusted models. All
significant Model 1 associations remain significant after adjust-
ment for Wave I and III socioeconomic, psychological and health
factors (Model 2) and after further adjustment for Wave IV de-
mographic and socioeconomic factors (Model 3).
VariableMean (s.e.) or %Range
High subjective relative debt
Household debts ($ thousands)
Calculated debt-to-asset ratio
Systolic blood pressure (mmHg)
Diastolic blood pressure (mmHg)
Household income (thousands $)
Household assets (thousands $)
Health insurance (months)
Correlations among Wave IV socioeconomic variables.
CorrelationsDebtsAssets IncomeEducation Calc.
Unstandardized beta coefficients and (95% confidence intervals) for the association
of debt with health outcomes.
Model 1Model 2Model 3
Total household debts
Calculated debt-to-asset ratio
?0.04 (?0.10, 0.00)
DBP0.01 (?0.02, 0.04)
Subjective relative debt
SBP 0.84 (?0.16, 1.85)
DBP1.21 (0.41, 2.01)**
?0.01 (?0.03, 0.01)0.01 (?0.00, 0.03)0.02 (0.01, 0.04)**
?0.00 (?0.02, 0.01) 0.01 (0.00, 0.03)*0.02 (0.01, 0.04)**
?0.01 (?0.01, ?0.00)**0.00 (?0.00, 0.01)0.01 (0.00, 0.01)*
?0.05 (?0.13, 0.02)
0.00 (?0.06, 0.06)
?0.01 (?0.08, 0.06) ?0.01 (?0.09, 0.06)
0.03 (?0.03, 0.10) 0.04 (?0.02, 0.11)
0.03 (0.02, 0.05)**0.02 (0.01, 0.03)**0.02 (0.00, 0.03)**
0.02 (0.01, 0.04)**0.01 (0.00, 0.03)*0.01 (0.00, 0.02)*
0.01 (0.00, 0.01)**0.00 (0.00, 0.01)*0.00 (0.00, 0.01)*
?0.01 (?0.06, 0.03) ?0.01 (?0.06, 0.03)
0.02 (?0.01, 0.05)0.02 (?0.01, 0.05)
1.15 (0.92, 1.38)**0.83 (0.64, 1.03)**0.55 (0.34, 0.77)**
0.72 (0.50, 0.94)**0.42 (0.23, 0.62)**0.33 (0.15, 0.53)**
0.24 (0.17, 0.31)**0.13 (0.08, 0.19)**0.10 (0.04, 0.16)**
0.83 (?0.13, 1.79)
1.15 (0.33, 1.98)**
0.71 (?0.32, 1.75)
1.01 (0.13, 1.91)*
Model 1: unadjusted association between debt and health outcomes.
Model 2: Model 1 þ all Wave I and Wave III confounders.
Model 3: Model 2 þ concurrent Wave IV socio-demographic factors.
E. Sweet et al. / Social Science & Medicine 91 (2013) 94e100
Author's personal copy
Our results show that, among young adults in the nationally
representative Add Health Study, household debt is a significant
independent predictor of health outcomes. This finding is an
important contribution to research on socioeconomic disparities in
health, since household debt is rapidly rising in this country and
studies exploring its general health impact are scarce. However, our
findings also suggest that how debt is operationalized matters for
its association with health; this speaks to its social patterning as
well as its complex nature as both a positive and negative socio-
While debt can be thought of as a negative socioeconomic
attribute, in the Add Health dataset absolute household debt is
positively correlated with other SES variables, such as income and
assets. This is not surprising when considering the possible ways in
which credit can function as a status attainment vehicle as well as
howcredit access is structured in this country. To obtain most types
of loans, recipients must typically demonstrate ability to meet loan
terms; credit scores, which are increasingly criticized as repre-
senting ambiguous and possibly biased calculations of credit
worthiness, are now used to determine eligibility for almost every
type of credit, from home mortgages to credit cards and even
hospital billing plans (Fremstad & Traub, 2011). In short, some level
of capital is typically necessary to accumulate large amounts of
debt, and thus debt would be expected to correlate with other as-
pects of SES. Furthermore, particularly in this young adult sample,
debt may be a vehicle for accumulating other forms of socioeco-
nomic status. For instance, high absolute levels of debt may reflect
student loans or other education-related debts, which while high
represent a beneficial personal socioeconomic investment. Or high
debt could reflect consumer spending that enhances symbolic
capital and is thus an investment in social prestige and status
In light of the potential of debt to reflect higher SES, it is perhaps
not surprising that absolute debt was associated with better self-
reported general health in unadjusted models. When other socio-
economic variables were included in the models, however, absolute
debt emerged a significant independent predictor of higher
perceived stress and depression and worse general health. This
suggests that absolute debt alone, particularly absent information
about how that debt is constituted, is an unreliable socioeconomic
determinant of health. It also suggests that debt has the most utility
as a health predictor when considered in relation to the broader
household financial context.
We considered the broader financial context in this study by
exploring debt in relation to another household socioeconomic
attribute e assets e with the expectation that debt will be most
problematic when it exceeds other household resources. Indeed we
found that high relative debt (debt-to-asset ratio), whether sub-
jectively assessed or calculated based on reported debt and asset
values, was associated with higher perceived stress and depression
and worse self-reported general health, even when accounting for
life-course health and economic conditions and other indices of
current socioeconomic position. The magnitude of the effect of high
relative debt in adjusted models was substantial; on average, in-
dividuals with a high compared to low subjective relative debt
reported 0.55 higher symptoms of perceived stress in adjusted
models (representing an 11.7% increase relative to the mean) and
0.33 higher depressive symptoms (a 13.2% increase relative to the
However, with respect to blood pressure, the operationalization
of the independent variable mattered. While calculated debt-to-
asset ratio was not associated with blood pressure, higher subjec-
tive relative debt was significantly associated with higher diastolic
blood pressure in both unadjusted and adjusted models (a 1.3%
increase in DBP relative to the mean in fully adjusted models).
While this effect is relatively small, it is clinically significant; a
2 mmHg increase in DBP, for instance, is associated with 17% higher
risk of hypertension and 15% higher risk of stroke (Cook, Cohen,
Hebert, Taylor & Hennekens, 1995). These findings indicate that
high relative debt is associated with worse psychological and
general health, but that with respect to blood pressure the psy-
chological feeling of being indebted may be more salient than
actual financial standing. We should note however, that our mea-
sure of subjective relative debt does not capture individual psy-
chological perceptions, such as attitudes towards debt and feelings
of indebtedness and financial strain, which maybe important tothe
way in which debt impacts health. Future research should aim to
better tease apart the effect of having debt from the perception of
being in debt and its associated emotional responses.
Limitations of the data could also influence interpretation of
the findings. In theory our subjective and calculated debt-to-asset
variables are indexing the same underlying condition, however
these variables were only 34% correlated. This could be for several
reasons. First, subjective assessment of one’s financial situation
may not accurately reflect the objective situation. Second,
reporting of absolute amounts of household debts and/or assets
may be inaccurate. Third, respondents were asked to exclude both
mortgage debt and home equity from reported absolute debt and
asset amounts, but to include home equity in their subjective
assessment of their net status. Property-related debts and assets
likely play a large role the financial situations of many households,
and thus the modest correlation between subjective and calcu-
lated debt-to-asset ratios and their differential associations with
some health outcomes could be due to this discrepancy in ques-
tion phrasing. Our current data does not allow us to disentangle
these possibilities, but future work should aim to clarify these
Furthermore, the Add Health cohort we examined is relatively
young (aged 24e32 years), and life course position has important
implications for debt. Our young sample, for instance, may not yet
have accumulated enough of a debt burden to allow us to see the
full impact of debt on health. Or, conversely, large levels of debt in
young adults may be less meaningful for health than among older
adults. The permanent income hypothesis (Friedman, 1957) and
related life-cycle theory of consumption (Modigliani & Brumberg,
1954) suggest that consumption decisions that may lead to debt
are based on anticipated long-term household economic prospects
more than immediate realities (i.e. the practice of ‘consumption
smoothing’ over the life course). Under this hypothesis, instances of
high debt, even relative to assets, in this young adult sample may
not truly index financial strain. The permanent income hypothesis,
however, has been poorly supported by evidence and criticized for
oversimplifying the complex realities and social contexts of
household consumption decisions (Alvarez-Cuadrado & Long, 2011;
Palley, 2010; Wisman, 2009). Furthermore, we found relative debt
to be predictive of health outcomes even after adjusting for early
life SES, which can be seen as a proxy for life course socioeconomic
expectations. This speaks to the apparent psychological potency of
debt despite rational economic justifications. This does not, how-
ever, diminish the need for future research to address the issue of
life course status and the position of debt within broader socio-
Along similar lines, while we explored three different indices of
debt in this paper, additional work is needed to elaborate and un-
derstand the impacts of diverse dimensions of debt on health. Prior
studies have focused separately on credit card debt, home fore-
closures, debt-related stress, and participation in credit counseling
(Drentea & Lavrakas, 2000; McLaughlin et al., 2011; Munster et al.,
E. Sweet et al. / Social Science & Medicine 91 (2013) 94e100
Author's personal copy
2009; O’Neill et al., 2006), but the comparative effects of these
different dimensions with respect to health outcomes are un-
known. Furthermore different types of debt may have different
material, psychological and social meanings, and may occur in
different contexts. Payday loans, for instance, have been singled out
as particularly socially stigmatized. These short-term, revolving
high-interest loans are more common in lower income and mi-
nority neighborhoods, where mainstream banking facilities have
been replaced with non-traditional lending outlets (Barr, 2004;
Logan & Weller, 2009; Williams, 2008). Payday loans are generally
considered highly predatoryand have been described as symbols of
“a devalued place occupied by devalued people” (Williams, 2005).
Thus, while the relative size of payday or other short-term loans
may be small compared to a home mortgage or student loan, their
psychosocial impact could be much greater. It will be important for
future research to account for these potential differences in types of
debt as well as the different sociocultural, political-economic, and
even national contexts in which they occur. While our discussion
and analyses are specific to the United States, debt is a global
phenomenon. The ways in which debt is differentially structured
and experienced in other regions and countries (Alfaro & Gallardo,
2012; Aniola & Golas, 2012) and its consequent impact on health
should also be examined.
Finally, future work should aim to specify the mechanisms
through which debt impacts health. Considering the established
evidence of the psychological impact of debt, mechanisms
involving physiological and behavioral consequences of chronic
stress are likelycandidates. Dysregulation of basal cortisol rhythms,
elevations in blood pressure and inflammatory markers, and
metabolic alterations associated with health behaviors are all
linked with chronic stress exposure and are important indicators of
disease susceptibility (McEwen, 1998, 2003). Identifying the rela-
tionship of debt to these biomarkers would shed valuable light on
the pathways through which debt becomes embodied. Further-
more, being in debt or perceiving oneself to be indebted may limit
access to health resources or impact health decision-making.
Indeed, one third of indebted households have been found to
regularly forego medical care in efforts to reduce family expenses
(Garcia & Draut, 2009), and two-thirds of individuals with medical
debt have reported changing their medical care seeking behavior as
a result of their debt (O’Toole, Arbelaez & Lawrence, 2004). Clearly,
this is another pathway through which debt can influence health
and should be explored to fully understand debt’s impact.
In this paper we have provided evidence that, in addition to
known associations with psychological health, financial debt is
associated with worse self-reported physical health and blood
pressure. In testing multiple indices of debt, we found high
household debts relative to assets to be the most consistent and
robust predictor of health outcomes. We also found that a high
subjective assessment of indebtedness was the strongest predictor
of blood pressure, suggesting that psychological dimensions of debt
may be particularly salient when it comes to cardiovascular health.
Importantly, we used longitudinal data to control for prior health
related and socio-demographic factors that could influence the
relationship of debt with health. Much additional work is needed,
though, to better elucidate the mechanisms through which debt
may impact health and to operationalize the types of debt that
matter most for health.
This research uses data from Add Health, a program project
directed byKathleen Mullan Harris and designed byJ. Richard Udry,
Peter S. Bearman, and Kathleen Mullan Harris at the University of
North Carolina at Chapel Hill, and funded by grant P01-HD31921
(gs1) from the Eunice Kennedy Shriver National Institute of Child
Health and Human Development, with cooperative funding
from23 otherfederal agencies
acknowledgment is due Ronald R. Rindfuss and Barbara Entwisle
for assistance in the original design. Information on how to
obtain the Add Health data files is available on the Add Health
website (http://www.cpc.unc.edu/addhealth). No direct support
was received from grant P01-HD31921 for this analysis. Support
for the analyses presented in this paper was provided by grant #
1R01 HD053731-01 (McDade) from the Eunice Kennedy Shriver
National Institute of Child Health and Human Development.
Adler, N. E. (2009). Health disparities through a psychological lens. American Psy-
Adler, N. E., Boyce, T., Chesney, M. A., Cohen, S., Folkman, S., Kahn, R. L., et al. (1994).
Socioeconomic status and health. The challenge of the gradient. American
Psychologist, 49, 15e24.
Adler, N. E., & Stewart, J. (2010). Health disparities across the lifespan: meaning,
methods, and mechanisms. Annals of the New York Academy of Sciences, 1186,
Alfaro, R., & Gallardo, N. (2012). The determinants of household debt default. Revista
de Analisis Economico, 27, 55e70.
Alvarez-Cuadrado, F., & Long, N. V. (2011). The relative income hypothesis. Journal of
Economic Dynamics and Control, 35, 1489e1501.
Aniola, P., & Golas, Z. (2012). Differences in the level and structure of household
indebtedness in the EU Countries. Contemporary Economics, 6, 46e59.
Barr, M. S. (2004). Banking the poor. Yale Journal on Regulation, 21, 123e128.
Braveman, P. A., Cubbin, C., Egerter, S., Chideya, S., Marchi, K. S., Metzler, M., et al.
(2005). Socioeconomic status in health research: one size does not fit all.
Journal of the American Medical Association, 294, 2879e2888.
Bridges, S., & Disney, R. (2010). Debt and depression. Journal of Health Economics, 29,
Brown, S., Taylor, K., & Price, S. W. (2005). Debt and distress: evaluating the psy-
chological cost of credit. Journal of Economic Psychology, 26, 642e663.
Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A global measure of perceived
stress. Journal of Health & Social Behavior, 24, 385e396.
Cook, N. R., Cohen, J., Hebert, P. R., Taylor, J. O., & Hennekens, C. H. (1995). Impli-
cations of small reductions in diastolic blood pressure for primary prevention.
Archives of Internal Medicine, 155, 701e709.
Dossey, L. (2007). Debt and health. Explore (NY), 3, 83e90.
Drentea, P., & Lavrakas, P. J. (2000). Over the limit: the association among health,
race and debt. Social Science & Medicine, 50, 517e529.
Drentea, P., & Reynolds, J. R. (2012). Neither a borrower nor a lender be: the relative
importance of debt and SES for mental health among older adults. Journal of
Aging and Health, 24, 673e695.
Dudley, K. M. (2000). Debt and dispossession: Farm loss in America’s heartland.
Chicago: University of Chicago Press.
Entzel, P., Whitsel, E. A., Richardson, A., Tabor, J., Hallquist, S., Hussey, J., et al. (2009).
Add health wave IV documentation: Cardiovascular and anthropometric measures.
Chapel Hill, NC: Carolina Population Center.
Fremstad, S., & Traub, A. (2011). Discrediting America: the urgent need to reform the
nation’s credit reporting industry. New York, NY: Demos.
Friedman, M. (1957). A theory of the consumption function. Princeton: Princeton
Garcia, J. (2007). Borrowing to make ends meet: the rapid growth of credit card debt in
America. New York: Demos.
Garcia, J., & Draut, T. (2009). The plastic safety net: How households are coping in a
fragile economy. New York: Demos.
Graeber, D. (2011). Debt: The first 5000 years. New York: Melville House.
Gruenstein Brocian, D., Wei, L., & Ernst, K. S. (2010). Foreclosures by race and ethnicity:
The demographics of a crisis. Washington, D.C.: Center for Responsible Lending.
Harvey, D. (2010). The enigma of capital and the crisis this time. American Socio-
logical Association Annual Meeting. Atlanta.
Jenkins, R., Bhugra, D., Bebbington, P., Brugha, T., Farrell, M., Coid, J., et al. (2008).
Debt, income and mental disorder in the general population. Psychological
Medicine, 38, 1485e1493.
Kim, J., Garman, E. T., & Sorhaindo, B. (2003). Relationships among credit counseling
clients’ financial well-being, financial behaviors, financial stress events, and
health. Financial Counseling and Planning, 14, 75e87.
Logan, A., & Weller, C. E. (2009). Who borrows from payday lenders? And analysis of
newly available data. Center for American Progress.
McEwen, B. S. (1998). Stress, adaptation, and disease. Allostasis and allostatic load.
Annals of the New York Academy of Sciences, 840, 33e44.
McEwen, B. S. (2003). Interacting mediators of allostasis and allostatic load: to-
wards an understanding of resilience in aging. Metabolism, 52, 10e16.
McEwen, B. S. (2004). Protection and damage from acute and chronic stress: allo-
stasis and allostatic overload and relevance to the pathophysiology of psychi-
atric disorders. Annals of the New York Academy of Sciences, 1032, 1e7.
E. Sweet et al. / Social Science & Medicine 91 (2013) 94e100
Author's personal copy Download full-text
McEwen, B. S., & Seeman, T. (1999). Protective and damaging effects of mediators of
stress. Elaborating and testing the concepts of allostasis and allostatic load.
Annals of the New York Academy of Sciences, 896, 30e47.
McLaughlin, K. A., Nandi, A., Keyes, K. M., Uddin, M., Aiello, A. E., Galea, S., et al.
(2011). Home foreclosure and risk of psychiatric morbidity during the recent
financial crisis. Psychological Medicine, 1e8.
Matthews, K. A., & Gallo, L. C. (2011). Psychological perspectives on pathways
linking socioeconomic status and physical health. Annual Review of Psychology,
Meltzer, H., Bebbington, P., Brugha, T., Jenkins, R., McManus, S., & Dennis, M. S.
(2011). Personal debt and suicidal ideation. Psychological Medicine, 41, 771e778.
Mitchell, J., & Jackson-Randall, M. (2012). Student loan debt tops $1 trillion. Wall
Modigliani, F., & Brumberg, R. (1954). Utility analysis and the consumption func-
tion: an interpretation of cross-sectional data. In K. K. Kurihara (Ed.), Post
Keynesian economics. New Brunswick: Rutgers University Press.
Munster, E., Ruger, H., Ochsmann, E., Letzel, S., & Toschke, A. M. (2009). Over-
indebtedness as a marker of socioeconomic status and its association with
obesity: a cross-sectional study. BMC Public Health, 9, 286.
O’Neill, B., Prawitz, A. D., Sorhaindo, B., Kim, J., & Garman, E. T. (2006). Financial
Distress/financial well-being for debt management program clients. Financial
Counseling and Planning, 17, 46e62.
O’Toole, T. P., Arbelaez, J. J., & Lawrence, R. S. (2004). Medical debt and aggressive
debt restitution practices: predatory billing among the urban poor. Journal of
General Internal Medicine, 19, 772e778.
Palley, T. I. (2010). The relative permanent income theory of consumption: a syn-
thetic KeyneseDuesenberryeFriedman model. Review of Political Economy, 22,
Pollack, C. E., & Lynch, J. (2009). Health status of people undergoing foreclosure in
the Philadelphia region. American Journal of Public Health, 99, 1833e1839.
Radloff, L. S. (1977). The CES-D Scale: a self-report depression scale for research in
the general population. Applied Psychological Measurement, 1, 385e401.
Reading, R., & Reynolds, S. (2001). Debt, social disadvantage and maternal
depression. Social Science & Medicine, 53, 441e453.
Selenko, E., & Batinic, B. (2011). Beyond debt. A moderator analysis of the rela-
tionship between perceived financial strain and mental health. Social Science &
Medicine, 73, 1725e1732.
Shavers, V. L. (2007). Measurement of socioeconomic status in health disparities
research. Journal of the National Medical Association, 99, 1013e1023.
StataCorp.. (2009). Stata statistical software: Release 11. College Station, TX: Stata-
Sweet, E. (2011). Symbolic capital, consumption, and health inequality. American
Journal of Public Health, 101, 260e264.
Williams, B. (2005). Debt for sale: a social history of the credit trap. Philadelphia:
University of Pennsylvania Press.
Williams, B. (2008). The precipice of debt. In J. Collins, M. di Leonardo, & B. Williams
(Eds.), New landscapes of inequality: Neoliberalism and the erosion of democracy
in America. Santa Fe: School for Advanced Research Press.
Wisman, J. D. (2009). Household saving, class identity, and conspicuous con-
sumption. Journal of Economic Issues, 43, 89e114.
E. Sweet et al. / Social Science & Medicine 91 (2013) 94e100