PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of geronbLink to Publisher's site
 
J Gerontol B Psychol Sci Soc Sci. Sep 2012; 67(5): 585–594.
Published online Aug 16, 2012. doi:  10.1093/geronb/gbs056
PMCID: PMC3536554
Editor's choice

Accumulated Financial Strain and Women’s Health Over Three Decades

Abstract

Objective.

Drawing from cumulative inequality theory, this research examines how accumulated financial strain affects women’s self-rated health in middle and later life.

Method.

Using data from the National Longitudinal Survey of Mature Women (1967–2003), we employ random-coefficient growth curve models to examine whether recurring financial strain influences women’s health, above and beyond several measures of objective social status. Predicted probabilities of poor health were estimated by the frequency of financial strain.

Results.

Financial strain is associated with rapid declines in women’s health during middle and later life, especially for those women who reported recurrent strain. Changes in household income and household wealth were also associated with women’s health but did not eliminate the effects due to accumulated financial strain.

Discussion.

Accumulated financial strain has long-term effects on women’s health during middle and later life. The findings demonstrate the importance of measuring life course exposure to stressors in studies of health trajectories.

Key Words: Cumulative advantage/disadvantage theory, Multi-level models, Economic hardship, Stress, Cumulative Inequality Theory, Life course.

Background

Research in gerontology and life course epidemiology is giving renewed attention to the long-term antecedents of health in later life. Central to this genre of research in recent years are efforts to assess life course exposure to stressors that may exact a toll on one’s health. All people face stressors at some time in their lives, as Longfellow’s adage conveys: “Into each life some rain must fall.” Life course studies of health, however, are calling attention to repeated exposure to stressors because of the growing body of evidence revealing that chronic strain is consequential for physical and mental health (Avison & Turner, 1988; Willson, Shuey, & Elder, 2007).

It is well established that financially disadvantaged individuals have poorer health than their advantaged counterparts on a variety of outcomes, ranging from self-rated health (Janzen & Muhajarine, 2003) to mortality (Szanton et al., 2008). Financial strain, as a form of stress proliferation, may also affect health through multiple pathways including the erosion of personal control and mastery, loss of supportive social relationship, family tension, and resource constraints (Kahn & Pearlin, 2006; Pearlin, Aneshensel, & LeBlank, 1997). Yet, most studies examining health status are cross-sectional in nature, relying on point estimates of income, financial strain, or subjective social class (Adler & Ostrove, 1999; Baron-Epel & Kaplan, 2009; Singh-Manoux, Adler, & Marmot, 2003). Recognizing that financial strain, operationalized as arising from economic insufficiency, may ebb and flow during adulthood, the purpose of this analysis is to assess the health consequences of accumulated strain. We anticipate that a single bout of financial strain will not be consequential to health, but that chronic financial stress will result in steeper health declines during adulthood.

Financial Strain and Health

This investigation of accumulated financial strain and health is informed by a large literature on the social gradient in health but more specifically by studies focused on the strain or stress associated with limited material resources. In surveying the current literature, there appear to be at least two ways for future research to improve our understanding of the relationship between financial strain and health.

First, there are relatively few studies of financial strain and health that explicitly address life course variation in strain. Among those that use longitudinal data, the time frame ranges from a 6-month to a 25-year follow-up, respectively (Macleod, Davey Smith, Metcalfe, & Hart, 2005; Operario, Adler, & Williams, 2004), and most but not all show that low socioeconomic standing or financial strain is associated with poorer health. Careful examination of the longitudinal studies, however, shows that most of them track only the health measures over time; financial strain or subjective social status is typically measured only once (Operario et al., 2004; Singh-Manoux, Marmot, & Adler, 2005). Inasmuch as strain ebbs and flows, using one measurement occasion may miss important variation over time. More importantly, single-occasion measures of financial strain preclude estimation of the recurrent nature of financial strain, which may be especially vexing to those who face such misfortune.

Second, most studies of financial strain rely on income and occupational status as the sole indicators of objective financial and social status. These indicators, however, may be less meaningful when studying middle and later life, especially among retired persons. It has long been recognized that wealth is a more meaningful and predictive indicator of material well-being because it reflects lifetime accumulation of finances and status (Henretta & Campbell, 1978). Limited income may be less of a concern for older people who have sizable financial reserves, in part because they seek to limit their income to preserve their capital. Surprisingly few studies, however, incorporate any measure of wealth, and even fewer account for changing wealth over the life course.

Despite the limitations of prior studies (e.g., cross-sectional, reliance on a single measure of financial status), we identified several articles from the past decade that point to the importance of considering life course variability in financial strain. Among the exemplars, one longitudinal study of Mexican American older adults included measures of financial strain and health measured at two time points, showing that financial strain is associated with multiple measures of poor health, even after accounting for household income (Angel, Frisco, Angel, & Chiriboga, 2003).

Three additional studies used retrospective questions from cross-sectional surveys to examine how long-term financial strain influences health. Kahn and Pearlin (2006, p. 24) also studied older people and found that “the greater the persistence of financial strains across the earlier years of the life course, the greater the damage to multiple dimensions of late-life health.” They found that women were more likely to have higher numbers of serious health conditions, functional impairments, and depressive symptoms compared with men. However, they did not discuss the effects of financial strain on health by gender. Cambois and Jusot (2010) studied French men and women 35 years or older and reported that lifelong adverse experiences, including financial and housing problems, were associated with multiple health outcomes even after controlling for education, occupation, and income. The authors found that the effects of adverse life events on health were particularly strong for women. Specifically, adverse life events (including financial strain) accounted for a quarter of the odds for poor self-rated health and over a third of the odds for chronic disease and activity limitations for women in the lower income quintiles. Szanton, Thorpe, and Whitfield (2010) showed that recurrent financial strain was associated with poor health among African Americans but that adult financial strain was more consequential than childhood financial strain. They did not test for gender differences in these effects. These studies are exemplary because they point to the value of incorporating measures tapping accumulated financial strain and suggest that single-occasion estimates of financial strain probably underestimate the effect of financial strain on health. At the same time, none of these studies consider wealth, and few examine the effects separately for women.

The lack of attention to studying the relationship between financial strain and health among women is important because financial strain may affect their well-being in ways that are distinct from men. To begin, women are overrepresented among those living in poverty and are paid less than men, even when they have the same level of education and are in the same occupational field (Gomel, 1997; U.S. Census Bureau, 2006). In middle age, women are more likely than men to be single heads of household and to carry the responsibility for raising children with fewer economic resources (U.S. Census Bureau, 2006). In later life, women have a higher risk of becoming poor and generally remain in that state longer after entering it (Cellini, McKernan, & Ratcliffe, 2008). Thus, chronic financial strain may be especially consequential to women’s health (Groh, 2007). Indeed, some previous research reveals that the effects of financial strain may be far-reaching on women’s mental (Kessler et al., 2003) and physical health (Du, Fang, & Meyer, 2008; Hemingway, 2007).

This study is designed to overcome several of the limitations of previous research by incorporating measures of wealth, financial strain, and health in a long-term prospective study of women.

Accumulation and Social Comparison Processes

This study draws from cumulative inequality (CI) theory to explore the life course processes whereby financial strain and objective socioeconomic status (SES) influence health (Ferraro & Shippee, 2009). CI theory is a middle-range one that synthesizes elements from stress process (Pearlin, Mullan, Semple, & Skaff, 1990), cumulative disadvantage (Dannefer, 2003), and life course perspectives (Elder, 1998) to emphasize the accumulation of inequality over the life course. For the present project, it is useful in at least three respects. First, the theory emphasizes how trajectories are shaped by exposures. Rather than assuming that early disadvantage (i.e., financial strain) results in health decline, the idea is that the accumulation of exposure to stressors (and resources) is pivotal to predicting health outcomes (Astone, Misra, & Lynch, 2007; Cambois & Jusot, 2010). Accordingly, one hypothesizes that the health consequences would be greatest for chronic financial strain. A single period of financial strain may well have no enduring effect on health.

Second, the theory emphasizes how trajectories in one domain may influence other domains. For this study, our focus is the relationship between financial and health domains—wealth and health. The theoretical expectation is for domain diffusion such that accumulated disadvantage in one domain spills over into other life domains, thereby shaping other life trajectories. Although we are limited in our ability to fully measure domain spillover, we recognize that both domains can mutually affect each other and therefore adjust for baseline measures of health. This is more than just studying the relationship between wealth and health at any point in time. Rather, recognizing that both domains are changing, the aim is to examine the association between variability in exposure to financial strain and one’s health trajectory.

Third, the theory emphasizes that how actors view their life trajectories is consequential to their well-being. Structural disadvantages across the life course may well lead to poorer health and quality of life, but people’s evaluations of their life situations can also be consequential for health (Ferraro & Shippee, 2009). Although objective circumstances are central to the accumulation of inequality, people continually evaluate their placement on status hierarchies. Indeed, social comparison theories show that individuals develop an awareness of their life circumstances, which may have consequences for health and well-being (Carstensen, 2006; Gilbert, Giesler, & Morris, 1995). There are studies of financial strain that assess it by reported income in relation to poverty thresholds in the United States (Lynch, Kaplan, & Shema, 1997) or Canada (Janzen & Muhajarine, 2003). Theoretically, however, financial strain also has a subjective element by which people compare their financial lot to that of their peers. In this way, financial strain is a psychosocial stressor, above and beyond reported financial resources, which has been found to influence health (Siegrist & Marmot, 2006; Sun, Hilgerman, Durkin, Allen, & Burgio, 2009).

To summarize, this study is designed to extend the previous literature in three main ways. First, we take advantage of longitudinal data from a nationally representative sample of women to prospectively study financial strain and health (without relying on retrospective reports of financial status). Second, unlike any other study of which we are aware, the analytic models presented below incorporate change in multiple measures of financial well-being: accumulated financial strain, household income, and household wealth. Finally, we investigate the potential influence of financial strain on health during a major portion of the adult life course, from middle age into later life and spanning 36 years. Although we are unable to account for childhood financial strain in this study, prior research reveals that adult financial strain is more consequential to adult health (Szanton et al., 2010). Two main hypotheses guide the analysis: (a) Financial strain has an independent effect on health, above and beyond the effects due to income, occupational prestige, and wealth. (b) Accumulated financial strain will result in steeper health declines during adulthood than financial strain measured once.

Methods

Sample

This study used data from the National Longitudinal Survey of Mature Women (NLSMW). Multistage probability sampling was used to draw a representative sample of civilian, non-institutionalized women aged 30–44 years in 1967, with an oversample of Black women. The original sample in 1967 contained 5,083 women. Since then, the women have been interviewed a total of 21 times through 2003, when 2,237 (44%) of the original respondents were surveyed. Most sample attrition was due to the death of respondents (N = 1,485; 29.2%), refusals (N = 1,036; 20%), and failure to locate (N = 325; 6%; U.S. Department of Labor, 2001).

Our final sample consisted of 3,181 women for whom valid information was available on the response variable. For most NLSMW measures, there is little missing data, but single imputation was used to address item-missing data on household wealth. Findings were robust to alternate strategies of handling missing data, including list-wide deletion.

Self-Rated Health

Data on women’s self-rated health were drawn from seven time points over the course of the study—1967, 1992, 1995, 1997, 1999, 2001, and 2003. Respondents were asked to rate their health in comparison to other women their age as “excellent,” “good,” “fair,” or “poor” (coded from 4 to 1, respectively). As described in more detail below, we preserved the ordinal nature of the variable for the growth curve analysis.

Financial Strain

Measures of financial strain were derived from the question “Which of these four statements best describes your (family’s) ability to get along on your (its) income?” Response categories include 1 “I (We) always have money left over” 2 “I (We) have enough with a little extra sometimes” 3 “I (We) have just enough, no more” and 4 “I (We) can’t make ends meet.” This measure is similar to those used in other studies tapping financial strain (Ferraro & Su, 1999; Szanton et al., 2008). We used data from the 1979, 1981, 1982, 1984, 1986, and 1987 waves of the study, and created a series of binary variables, with 1 equal to “can’t make ends meet” and 0 equal to all else. We then created a count variable of financial strain ranging from 0 to 4, with 0 equal to no financial strain and 4 equal to chronic financial strain. In our analyses, we top-coded financial strain at 4 because this category was substantively meaningful and represented respondents who experienced financial strain the majority of the time. Alternative ways of coding financial strain yielded similar results. In addition, we modeled financial strain as a time-constant effect because lagging the financial variables (through 1987) helps reduce the risk of endogeneity. Substantively, however, the literature shows that financial strain is likely to be chronic (Kahn & Pearlin, 2006) and it often has lasting effects on physical and mental health (Lynch et al., 1997; Price, Choi, & Vinokur, 2002).

Objective Measures of Financial Well-Being

Objective measures of financial well-being were assessed using two main indicators: household income and household wealth, each operationalized as a time-varying covariate (TVC). Data for household income were drawn from six different waves (1992–2003) and coincide with reports of self-rated health. The household income variable is a sum of any and all income reported by household members (including the woman’s partner, if applicable).

We used summary variables of wealth at six different waves to represent respondents’ total net family assets from 1989 to 2003. This variable was created by adding the individual’s housing, savings, bonds, IRA, insurance, and business assets; and then subtracting mortgages, loans, and any other debts (excluding automobile value). Wealth was not available in 1992 and was therefore substituted with household wealth in 1989. In addition to these measures of financial well-being, we also account for women’s occupational prestige and education. Prestige scores (two-digit codes) have a theoretical range of 0–97 and are based on income and education distributions associated with occupations identified in the 1960 Census (U.S. Department of Labor, 2001). Prestige scores were assigned to respondents’ current or most recent occupation. We included a baseline measure and a change score for occupational prestige from 1967 to 1977 (the period when most women went from around age 40 to approximately 54). Education was measured in years of formal education attended by 1967.

Covariates

Based on the literature examining financial strain and health, we adjusted for several additional variables. Chronological age is measured in years. We employed a Heckman selection model and included the variable expressing the mortality hazard in our growth curve analysis (Heckman, 1979). This covariate adjusts the estimates for sample selection due to mortality. Black is a binary variable coded 1 for Black respondents, and coded 0 for non-Black respondents (hereafter, referred to as White).

Locus of control is a measure ranging from 1 to 4, with higher values corresponding to a greater sense of personal control. This measure is based on the 11-item version of Rotter’s (1966) Internal-External Control Scale; values represent the mean score on items (α = 0.65). Similarly, depressive symptoms is measured using the 20-item Center for Epidemiologic Studies Depression scale, with values equal to the mean score (α = 0.89). Each item measure ranges from 1 “rarely or none of the time” to 4 “most or all of the time;” however, in creating the scale of depressive symptoms, we computed the row mean which retains the 1–4 scale. Given the lengthy gap in the measurement of self-rated health during the first half of the study and in an effort to account for endogeneity of health problems, we also adjust for self-rated health at baseline, coded as excellent or good v. fair or poor (while using the last six measures in ordinal form to estimate a growth curve). Marital status was measured as a TVC, with a series of binary TVCs representing married, widowed, divorced/separated, and never married (married served as the reference group). “Children” is a count variable representing the number of children born to respondents by 1977 when women were aged 40–55. These data provide a near complete picture of women’s child-bearing history (U.S. Department of Labor, 2001).

Analytic Strategy

To evaluate the relationship between financial strain and women’s self-rated health, we estimated a series of nested growth curve models. The data array over 36 years permitted the use of growth curve models to examine variations in women’s health trajectories. This particular case of multilevel modeling observes change in the dependent variable and allows us to partition the variability in self-rated health into two components: between-person variability and within-person variability (Singer & Willett, 2003). The former expresses the amount of variation in self-rated health due to differences between persons, whereas the latter expresses the variability in self-rated health within persons over six measurement occasions. For model specification, self-rated health was measured between 1992 and 2003. Financial strain preceded the measurement span of our dependent variable and thus was treated as a time-constant variable (between-subjects). In contrast, household income and household wealth were measured as time-varying (within-subjects). The analysis was completed using the Stata program gllamm for ordered responses; adaptive quadrature estimation was used to provide more accurate estimates (Rabe-Hesketh & Skrondal, 2008). The modeling features two levels of analysis, with occasions (level-1) nested within subjects (level-2). Mean-centered age was used as the time metric, along with a quadratic term to account for possible non-linearity in women’s health trajectories.

The analysis was divided into two main stages. First, random-intercept and random-coefficient models were used to estimate the variability about the mean intercept and slope and examine the effects of financial strain and more objective measures of SES on women’s self-rated health. The first three models are random-intercept models and allow the intercept of self-rated health to vary across women. Model 1 estimated the effect of demographic and psychological resources on women’s self-rated health, and Model 2 examined the effect of financial strain net of these covariates. Model 3 accounts for women’s changing income and wealth. The final model is a random-coefficient model that includes an interaction term between our time metric (mean-centered age) and women’s financial strain in order to predict the variability in women’s rates of change in self-rated health over time.

Second, we examined the role of accumulated financial strain as a predictor of changing self-rated health. To do so, we first estimated a random-coefficient model using number of occasions of financial strain as predictors. We used postestimation predicted probabilities that were based on the fixed effects derived from an adaptation of Model 3, where financial strain was treated as a set of binary variables (reference = no financial strain). The probabilities can be read as the percent chance of favorable reports of self-rated health (e.g., a .20 probability equals a 20% chance). Figure 1 presents predicted probabilities of self-rated health by number of reports of financial strain, adjusting for the full vector of covariates.

Figure 1.
Predicted probabilities of self-rated health by reports of financial strain.

Results

As shown in Table 1, there is a slight monotonic decline in self-rated health over time from 1992 to 2003. Women reported a mean of 0.71 occasions of financial strain, with one third of all women reporting at least one occasion of financial strain. The average woman reported an annual household income of $26,751 and the mean household wealth was $135,127. Most women were married, with 11 years of education on average.

Table 1.
Means and Standard Deviations of Variables in the National Longitudinal Survey of Mature Women, 1967–2003 (N = 3,181)

In Table 2, we display the final results from the growth curve modeling to examine whether financial strain (specified as a count variable, ranging from zero to four) is consequential to health during the 36-year observation period. We present coefficients with asterisks indicating level of significance (and standard errors), followed by odds ratios in the next column. Three nested models in Table 2 present (a) demographic and psychosocial resources only, (b) add financial strain, and (c) add household income and wealth. The intercept-only model is the simplest model (not shown), providing estimates of the proportion of between-person variability in women’s health trajectories. The random-intercept variance was estimated at 7.595, with a standard error of 0.287. This implies an intraclass correlation of 0.70 and indicates that 70% of the variability in self-rated health is due to differences between persons, not within.

Table 2.
Proportional-Odds Growth Curve Models for Self-Rated Health in the National Longitudinal Survey of Mature Women, 1992–2003

Model 1 includes the time-metric (mean-centered age) and estimates the effect of demographic and psychosocial resources on self-rated health using a random-intercept model. Both the linear and quadratic terms for age were negatively associated with self-rated health, indicating nonlinear health trajectories, with older adults reporting lower self-ratings of health on average. Mortality selection was a significant and negative predictor of self-rated health in Model 1, indicating that there is some selective mortality in the sample. Black women and those with depressive symptoms also reported lower self-rated health, whereas greater education, higher occupational prestige, and better self-rated health at baseline were significant and positive predictors of self-rated health. The random-intercept variance, which represents differences from the overall mean level of the response, was estimated at 5.823, with p < .001.

Model 2 includes financial strain along with other covariates. Financial strain was a significant, negative predictor of self-rated health: for every unit increase in financial strain, the odds of being in a higher category of self-rated health decreased by 28% (p < .001). Mortality selection was no longer significant. The random-intercept variance, after controlling for covariates, was estimated at 5.708.

Model 3 adjusts for time-varying measures of household income and wealth. Both household income and household wealth had a significant, positive effect on self-rated health, albeit the effect of wealth was small. The effects of financial strain attenuated slightly but remained consistent in its effects on women’s health when income and wealth were added to the model (p < .001). In other respects, results were generally similar to those found in Model 2. The random-intercept variance was equal to 5.554.

Finally, Model 4 adjusts for all covariates and introduces random slopes using an interaction term between the time metric and financial strain. The random-coefficient model allows us to estimate and predict the variability in slope in women’s health trajectories. After including the interaction and controlling for all other predictors, the effect of financial strain remained almost unchanged. The effects of household income and wealth also remained significant. The interaction term for financial strain was significant, indicating a significant effect on individual slopes over time for those reporting financial strain. Specifically, those with fewer occasions of financial strain had higher initial levels of self-rated health but experienced a steeper decline in self-rated health with age compared with their counterparts (possibly due to a ceiling effect). The estimated slope variance was 0.024 (p < .001), indicating significant between-person variation in slopes. Random-intercept variance was 6.211 (p < 0.001), also showing significant inter-individual variation in initial levels of self-rated health. The estimated covariance between the intercept and slope was 0.047, with the corresponding correlation estimated at 0.122. The positive covariance indicates that women with higher levels of self-rated health at baseline had smaller changes in self-rated health over time compared with those with low initial levels of self-rated health. A likelihood ratio test of Model 4 in comparison to the random-intercept model showed that the random-coefficient model had significantly better model fit.

We also tested for the effect of alternative cohort specifications on the trajectory of self-rated health but found no significant effects for cohort. For instance, including a binary cohort variable in the full model (1 = born between 1923 and 1930; 0 = born between 1931 and 1937) revealed a non-significant relationship with health.

Figure 1 illustrates the importance of accumulated financial strain for predicting self-rated health. We coded financial strain into a set of binary variables for displaying differences graphically. Women who reported one occasion of financial strain had a predicted probability of .17 to be in the “poor” category of self-rated health. Compare this to a predicted probability of .19 for those who reported two occasions of financial strain, .21 for those who reported three occasions of financial strain, and .32 for those who reported four or more occasions of financial strain. On the other hand, the probability of reporting excellent health dropped from .15 for women who experienced one occasion of financial strain to .06 for those with four or more occasions of financial strain.

Discussion

This study was guided by two hypotheses related to the relationship between financial strain and health. Unaware of any studies that considered financial strain along with measures of both income and wealth, our first hypothesis was designed to determine if there is an independent effect of financial strain on health after implementing extensive controls for financial status over time. The hypothesis was supported, revealing that financial strain, as a measure of psychosocial stress above and beyond reported financial resources, is related to health. In this sense, this study lends support to prior investigations of financial strain and health (Angel et al., 2003; Kahn & Pearlin, 2006). Even though time-varying covariates for income and wealth were accounted for in the prediction of health trajectories, financial strain remained an important influence (Sun et al., 2009). The findings are novel and point to the importance of recognizing subjective evaluations of financial strain for women’s health trajectories, even when controlling for objective indicators of health. They also provide support for cumulative inequality theory regarding the importance of subjective evaluations of trajectories. Women’s life course patterns place them at distinct risk for financial strain due to their position in the labor force, family roles, and lower earnings compared with men (U.S. Census Bureau, 2006) and this study gives priority to examining the relationship between financial strain and health among women.

It is important to place this conclusion in historical context. The period when financial strain was measured, 1979–1987, was a time of considerable economic change. Inflation was 11.3% in 1979 (13.5% in 1980) but much lower during the rest of the time period. Unemployment, by contrast, was less than 6% in 1979 but exceeded 9% in 1982–1983 before dropping to 6.2 in 1987 (U.S. Department of Labor, 2009). Thus, economic conditions changed notably during the period of observation, and different elements of the economy manifested problems during this time. Through it all, women who experienced chronic financial strain were more likely than women who experienced no or little financial strain to experience health declines.

Findings from this study differ from others in that our objective measures of SES also remained significant (cf. Matthews, Smith, Hancock, Jagger, & Spiers, 2005). This discrepancy could be due in part to differences in study design: many other studies on this topic used cross-sectional samples of older adults, usually with a single measure of objective income (Janzen & Muhajarine, 2003; Litwin & Sapir, 2009) or did not account for wealth (Matthews et al., 2005).

Objective measures of SES could impact health through inability to pay for basic services, providing safe living environment, access to health care, and other more direct effects (Adler & Ostrove, 1999). Yet, a significant part of the variation in health is not directly explained by economic aspects of SES, but by worry, anxiety, and depressive symptoms that often accompany economic strain (Hagquist, 1998; Lever, Pinol, & Uralde, 2005). We conclude that the subjective measurement of financial strain is important because it captures the psychosocial stress associated with material resources, rather than unilaterally assuming equivalent meanings for given levels of financial resources. Context matters, and asking the subject to evaluate her finances is an effective way to gage how she finds the adequacy of her finances for the context (Ennis, Hobfoll, & Schroder, 2000).

Most previous research has studied the effects of objective SES on individual’s health. Yet, not everyone with low income has the same objective experience. Some individuals have great difficulty getting by, while others may be able to tap into other resources such as social support, self-efficacy, and access to social services. Therefore, it is important to address both objective and subjective components of SES in their effects on health. In doing so, it is crucial to use robust measures of SES (and preferably more than one), and to utilize longitudinal, nationally representative data to better understand the mechanisms that operate in the relationship between SES (objective and subjective) and health.

Testing the second hypothesis focused on whether accumulated financial strain leads to more rapid declines in health declines. In all specifications tested, there was evidence that the recurrent nature of strain was consequential to health. Although most studies that examined accumulated financial strain relied on retrospective questioning, our results based on a prospective study lend support to earlier findings that strain is consequential to health (Cambois & Jusot, 2010; Kahn & Pearlin, 2006; Szanton et al., 2010). The findings presented herein also support those reported on other samples, such as older Hispanic Americans (Angel et al., 2003). One limitation of the current investigation is that our conclusions are limited to women. At the same time, we are unaware of any comparable study that tracks financial strain and health over such a substantial portion of the adult life course. By studying these women from middle age until they reached ages 66–80, one is able to observe the long-term consequences of chronic financial strain. Women who repeatedly reported financial strain were also more likely to report declining health over the 36-year period.

Prior studies have examined the effects of duration of objective measures of SES on health (Lynch et al., 1997; Power, Manor, & Matthews, 1999), but we are unaware of any study that has examined the relationship between recurrent financial strain and health while also incorporating wealth data from a nationally representative sample.

The findings are generally consistent with cumulative inequality theory regarding the importance of perceived trajectories and the role of accumulated exposures to influence health trajectories (Ferraro & Shippee, 2009). Specifically, we found strong cumulative effects of self-reported financial strain on self-rated health: the risk of poor health varied from 17% to 32%, depending on number of occasions of reported financial strain. Women who reported financial strain at four or more surveys, which we view as a situation of absorbing financial strain, had lower odds of being in a favorable category of self-rated health—a predicted probability of .06. Based on these analyses, the perception that one cannot escape financial strain may lead to a cascade of health problems.

We acknowledge some limitations of this research. As noted earlier, this research examined financial strain among women only. We accepted this restriction to be able to examine the phenomena under investigation over such a lengthy period of the life course. We welcome other studies that use both men and women to confirm or refute the conclusions presented herein. Second, due to data constraints, we modeled trajectories of self-rated health because they were available throughout the 36-year study. The NLSMW has limited measures of health that are consistently asked over the duration of the study. We do not presume that financial strain will have identical effects on other health conditions, especially ones that are diagnosed rather than self-reported. However, other studies (albeit cross-sectional), show that financial strain predicts multiple health outcomes (Angel et al., 2003; Szanton et al., 2010). Additional research on financial strain using different measures and outcomes is warranted. Finally, due to data limitations with time ordering of our key variables (the measurement of self-rated health is intermittent until the last decade of the study, and financial strain was asked only between 1979 and 1987), we were not able to fully test for domain spillover. It is possible that the endogeneity of financial strain may mean that poor health has induced financial strain. However, we account for self-rated health at baseline, which provides some evidence that the effect of financial strain on the trajectory of self-rated health is not solely due to initial poor health.

In conclusion, women beset by financial strain are at risk for more rapid health declines, above and beyond the effects due to actual differences in income and wealth. The stress of financial strain exacts an independent toll on women’s health. The most poignant finding, however, centered on the health consequences experienced by women who faced long, uninterrupted spells of financial strain. Based on the findings presented herein, eliminating health disparities must address the persistence of cumulative inequality in material resources. These results imply that even temporary relief from financial strain is beneficial to women’s health.

Funding

Support for this research was provided by the Fessler-Lampert Chair on Aging, University of Minnesota Center on Aging, and National Center for Research Resources of the National Institutes of Health to the University of Minnesota Clinical and Translational Science Institute (1UL1RR033183) to Dr. Shippee. Support was provided by the National Institute on Aging (T32AG025671) and the Purdue University Center on Aging and the Life Course to L. R. Wilkinson and K. F. Ferraro.

Acknowledgments

Manuscript prepared for presentation at the 2010 annual meeting of the Gerontological Society of America. The data were made available by the Inter-university Consortium for Political and Social Research, Ann Arbor, MI. Neither the collector of the original data nor the Consortium bears any responsibility for the analyses or interpretations presented here.

References

  • Adler N. E., Ostrove J. M. (1999). Socioeconomic status and health: What we know and what we don’t Annals of the New York Academy of Sciences 896 3–15. [PubMed]
  • Angel R. J., Frisco M., Angel J. L., Chiriboga D. A. (2003). Financial strain and health among elderly Mexican-origin individuals Journal of Health and Social Behavior 44 536–551. [PubMed]
  • Astone N. M., Misra D., Lynch C. (2007). The effect of maternal socio-economic status throughout the lifespan on infant birthweight Paediatrc and Perinatal Epidemiology 21 310–318. [PubMed]
  • Avison W. R., Turner R. J. (1988). Stressful life events and depressive symptoms: Disaggregating the effects of acute stressors and chronic strains Journal of Health and Social Behavior 29(3), 253–264. [PubMed]
  • Baron-Epel O., Kaplan G. (2009). Can subjective and objective socioeconomic status explain minority health disparities in Israel? Social Science & Medicine 69 1460–1467. [PubMed]
  • Cambois E., Jusot F. (2010). Contribution of lifelong adverse experiences to social health inequalities: Findings from a population survey in France European Journal of Public Health. Epub ahead of publication. [PubMed]
  • Carstensen L. L. (2006). The influence of a sense of time on human development Science 312 1913–1915. [PMC free article] [PubMed]
  • Cellini S. R., McKernan S.-M., Ratcliffe C. (2008). The dynamics of poverty in the United States: A review of data, methods, and findings Journal of Policy Analysis and Management 27 577–605.
  • Dannefer D. (2003). Cumulative advantage/disadvantage and the life course: Cross-fertilizing age and social science theory Journal of Gerontology, Social Sciences 58 S327–S337. [PubMed]
  • Du X. L., Fang S., Meyer T. E. (2008). Impact of treatment and socioeconomic status on racial disparities in survival among older women with breast cancer American Journal of Clinical Oncology 31 125–132. [PubMed]
  • Elder G. H., Jr. (1998). The life course as developmental theory Child Development 69 1–12. [PubMed]
  • Ennis N. E., Hobfoll S. E., Schroder K. E. (2000). Money doesn’t talk, it swears: How economic stress and resistance resources impact inner-city women’s depressive mood American Journal of Community Psychology 28 149–173. [PubMed]
  • Ferraro K. F., Shippee T. P. (2009). Aging and cumulative inequality: How does inequality get under the skin? The Gerontologist 49 333–343. [PMC free article] [PubMed]
  • Ferraro K. F., Su Y. (1999). Financial strain, social relations, and psychological distress among older people: A cross-cultural analysis Journal of Gerontology: Social Sciences 54B S3–S15. [PubMed]
  • Gilbert D. T., Giesler R. B., Morris K. A. (1995). When comparisons arise Journal of Personality and Social Psychology 69 227–236. [PubMed]
  • Gomel M. K. (1997). A focus on women Geneva: World Health Organization, Nations for Mental Health, Division of Mental Health and Prevention of Substance Abuse;
  • Groh C. J. (2007). Poverty, mental health, and women: Implications for psychiatric nurses in primary care settings Journal of the American Psychiatric Nurses Association 13 267–274.
  • Hagquist C. E. (1998). Economic stress and perceived health among adolescents in Sweden Journal of Adolescent Health 22 250–257. [PubMed]
  • Heckman J. (1979). Sample selection bias as a specification error Econometrica 47 153–161.
  • Hemingway A. (2007). Determinants of coronary heart disease risk for women on a low income: Literature review Journal of Advanced Nursing 60 359–367. [PubMed]
  • Henretta J. C., Campbell R. T. (1978). Net worth as an aspect of status American Journal of Sociology 83 1204–1223.
  • Janzen B. L., Muhajarine N. (2003). Social role occupancy, gender, income adequacy, life stage and health: A longitudinal study of employed Canadian men and women Social Science & Medicine 57 1491–1503. [PubMed]
  • Kahn J. R., Pearlin L. I. (2006). Financial strain over the life course and health among older adults Journal of Health and Social Behavior 47 17–31. [PubMed]
  • Kessler R. C., Berglund P., Demler O., Jin R., Koretz D., Merikangas K. R., Wang P. S. (2003). The epidemiology of major depressive disorder: Results from the National Comorbidity Survey Replication (NCS-R) Journal of the American Medical Association 289 3095–3105. [PubMed]
  • Lever J., Pinol N., Uralde J. (2005). Poverty, psychological resources and subjective well-being Social Indicators Research 73 375–408.
  • Litwin H., Sapir E. V. (2009). Perceived income adequacy among older adults in 12 countries: Findings from the Survey of Health, Ageing, and Retirement in Europe The Gerontologist 49 397–406. [PMC free article] [PubMed]
  • Lynch J. W., Kaplan G. A., Shema S. J. (1997). Cumulative impact of sustained economic hardship on physical, cognitive, psychological, and social functioning New England Journal of Medicine 337 1889–1895. [PubMed]
  • Macleod J., Davey Smith G., Metcalfe C., Hart C. (2005). Is subjective social status a more important determinant of health than objective social status? Evidence from a prospective observational study of Scottish men Social Science & Medicine 61 1916–1929. [PubMed]
  • Matthews R. J., Smith L. K., Hancock R. M., Jagger C., Spiers N. A. (2005). Socioeconomic factors associated with the onset of disability in older age: A longitudinal study of people aged 75 years and over Social Science & Medicine 61 1567–1575. [PubMed]
  • Operario D., Adler N. E., Williams D. R. (2004). Subjective social status: Reliability and predictive utility for global health Psychology and Health 19 237–246.
  • Pearlin L. I., Aneshensel C. S., LeBlank A. J. (1997). The forms and mechanisms of stress proliferation: The case of AIDS caregivers Journal of Health and Social Behavior 38 223–236. [PubMed]
  • Pearlin L. I., Mullan J. T., Semple S. J., Skaff M. M. (1990). Caregiving and the stress process: An overview of concepts and their measures The Gerontologist 30 583–594. [PubMed]
  • Power C., Manor O., Matthews S. (1999). The duration and timing of exposure: Effects of socioeconomic environment on adult health American Journal of Public Health 89 1059–1065. [PubMed]
  • Price R. H., Choi J. N., Vinokur A. D. (2002). Links in the chain of adversity following job loss: How financial strain and loss of personal control lead to depression, impaired functioning, and poor health Journal of Occupational Health Psychology 7 302–312. [PubMed]
  • Rabe-Hesketh S., Skrondal A. (2008). Multilevel and longitudinal modeling using Stata College Station, TX: Stata Press;
  • Rotter J. B. (1966). Generalized expectancies for internal versus external control of reinforcements Psychological Monographs 80 1–28. [PubMed]
  • Siegrist J., Marmot M. G. (2006). Social inequalities in health: New evidence and policy implications Oxford [UK]; New York: Oxford University Press;
  • Singer J. D., Willett J. B. (2003). Applied longitudinal data analysis: Modeling change and event occurrence (1st ed.). New York, NY: Oxford University Press, Inc;
  • Singh-Manoux A., Adler N. E., Marmot M. G. (2003). Subjective social status: Its determinants and its association with measures of ill-health in the Whitehall II study Social Science & Medicine 56 1321–1333. [PubMed]
  • Singh-Manoux A. Marmot M. G. Adler N. E. (2005). Does subjective social status predict health and change in health status better than objective status? Psychosomatic Medicine 67 855–861. [PubMed]
  • Sun F., Hilgerman M. M., Durkin D. W., Allen R. S., Burgio L. D. (2009). Perceived income inadequacy as a predictor of psychological distress in Alzheimer’s caregivers Psychology and Aging 24 177–183. [PMC free article] [PubMed]
  • Szanton S. L., Allen J. K., Thorpe R. J., Jr., Seeman T., Bandeen-Roche K., Fried L. P. (2008). Effect of financial strain on mortality in community-dwelling older women Journals of Gerontology: Social Sciences 63 S369–S374. [PMC free article] [PubMed]
  • Szanton S. L., Thorpe R. J., Whitfield K. (2010). Life-course financial strain and health in African-Americans Social Science & Medicine 71 259–265. [PMC free article] [PubMed]
  • U.S. Census Bureau (2006). American Community Survey: Selected economic characteristics Retrieved from http://www.census.gov/acs/www/index.html.
  • U.S. Department of Labor (2001). NLS of mature women user’s guide Washington, DC: Bureau of Labor Statistics and Center for Human Resource Research;
  • U.S. Department of Labor (2009). Consumer price index Washington, DC: Bureau of Labor Statistics;
  • Willson A. E., Shuey K. M., Elder G. H., Jr. (2007). Cumulative advantage processes as mechanisms of inequality in life course health American Journal of Sociology 112 1886–1924.

Articles from The Journals of Gerontology Series B: Psychological Sciences and Social Sciences are provided here courtesy of Oxford University Press