HRS Data Research Articles

No Rest for the Weary: Measuring the Changing Distribution of Retirement Wealth in the United States

ABSTRACT Since 1992 wealth for the bottom 90 percent of households nearing retirement has fallen. The only source of wealth helping the bottom 90 percent is Social Security. Despite pro savings policies and generous tax breaks for savings, the share of the bottom 50 percent having any retirement account didn’t change in 20 years — 46 percent in 1992 and 47 percent in 2016. Even the middle class suffered; the share of the next 40 percent with retirement savings fell from 85 percent in 1992 to a low of 71 percent in 2016. Housing ownership increased a bit for the bottom 50 percent but fell among the middle class and upper middle class. Home equity for the working and middle class fell. Using SCF and HRS data over 20 years, we find the bulk of working-class wealth is government social insurance. Economists should not exclude social insurance from wealth calculations. We find social insurance is the most important source of wealth for most families. Government policies and institutions have failed wealth building for most American households with workers.

Income-Related Inequalities in Physical and Cognitive Health Domains Over the Later Life Course: Longitudinal Evidence From the U.S. (1992-2016).

This study aims to investigate changes in the income-health gradient over the later life course. We test the age-as-leveler, the cumulative advantage/disadvantage, and the persistent inequality pattern for physical and cognitive health domains, and analyze whether these patterns are gendered. We used HRS data (1992-2016) and Poisson growth curve models to predict multimorbidity (33,860 participants) as an indicator of physical health and memory (25,291 participants) as an indicator of cognitive health. We disentangled the within-participant from the between-participant effects. For multimorbidity, the income-health gradient weakened as individuals aged; whereas for memory, the income-health gradient strengthened as individuals aged. The cumulative advantage/disadvantage of higher/lower income on memory may be more pronounced among women than men. Findings were confirmed by sensitivity analyses. Findings suggest that the support for the age-as-leveler or cumulative advantage/disadvantage pattern may depend on health domains and the effect strength may depend on gender.

US RACIAL AND ETHNIC DISPARITIES IN DEMENTIA LIFE EXPECTANCY, RISK/PROTECTIVE FACTORS, AND CAREGIVING STRESSORS

Abstract The health, emotional, and economic impacts of dementia are far-reaching and impossible to fully calculate. Late-life dementia is not equitably distributed across population subgroups including older people and racial/ethnic minorities. Clear evidence underscores disparities in social/personal resources, environmental exposures, racism/discrimination, increased burden of caregiving, and uneven access to quality care especially for Blacks and Hispanics and caregivers. Each paper in this symposium addresses high-priority, disparity research areas highlighted in the recent NASEM Decadal Survey of Behavioral and Social Science Research on ADRD: characteristics of population groups that influence disease ascertainment and dementia life expectancy; quantification of mid- and late-life risk/protective factors; pesticide exposure and cognition, and caregiving stressors among racial/ethnic minorities. All presentations utilize nationally representative data on older people in the US. Garcia et al. utilize HRS data to create dementia life expectancy estimates using four competing algorithmic classifications of dementia status by race/ethnicity (Blacks, Whites), nativity status (US-, non-US-born Latinos), and gender. Applying the Lancet Commission Dementia Prevention framework to HRS data, Ferdows and Aranda examine risk/protective factors and quantify the relative contribution of individual variables in explaining racial/ethnic dementia disparities. Chanti-Ketterl et al. elucidate the association between organochlorines pesticides (total scores, specific type) and cognition based on NHANES data to ascertain racial/ethnic differences. Based on NHATS, Moon et al. explore differences in primary stressors by dementia care and race (Blacks, Whites) during COVID-19. Findings provide evidence of potential multi-level targets for interventions and public policies. All primary authors are early-career scholars dedicated to mitigating ADRD health disparities.

BEREAVEMENT ACROSS THE ADULT LIFE COURSE: VARIATIONS IN EFFECTS ON HEALTH AND WELL-BEING

Abstract This symposium considers the consequences of deaths of family members across the life course on psychological well-being, physical health, and mortality in the middle and later years. The first two papers focus on intergenerational losses, using data from the HRS. Donnelly, Lin and Umberson examine the ways in which the impact of parental death on adult children’s health in later life are shaped by children’s ages at loss, social isolation, and race/ethnicity. Next, Huo and Kim investigate the extent to which the effect of child loss on parents’ well-being in mid- and later-life is moderated by volunteering. The third and fourth papers focus on the role of spousal loss in health and well-being. Rodin and colleagues use HRS data to explore whether the association between widowhood and mortality differs depending on the cognitive health of the surviving spouse. Next, Vedder and colleagues provide a systematic review of variations in patterns and predictors of loneliness by marital status in middle and later life. Finally, Miles and Olsen examine the association between recent loss and binge drinking among spouses and adult children who had served as caregivers to the deceased, using data from the Georgia Behavioral Risk Factor Surveillance Survey. This diverse set of studies reveals how the impact of the death of family members on survivors’ health and well-being is shaped by a complex set of structural and socioemotional factors, shedding new and important light on health disparities by sex/gender and race/ethnicity in the middle and later years.

Does Relationship End Precede Cognitive Decline? An Analysis of the Health and Retirement Study

Relationship status is thought to be associated with cognitive health in older adults, with married persons performing better on memory assessments than unmarried-cohabitating, single, divorced, and widowed persons. However, questions remain about whether relationship termination causes cognitive decline, is a result of it, or whether they share a cause; and the mechanisms by which such a relationship might operate. To address this gap in the literature, we hypothesized that relationship termination could affect cognition via the following five pathways: (1) post-termination depression; (2) loss of distributed-cognition partner; (3) cognitive depletion from caring for partner in declining and ultimately terminal health; (4) divorce to preserve assets to qualify for Medicaid to cover healthcare for cognitive decline; and (5) post-termination changes in neuropsychiatric symptoms alongside a pre-existing neurodegenerative condition that also causes cognitive decline. Using data from the 2000–2016 waves of the Health and Retirement Study (HRS; N = 23,393), we found that relationship termination, whether due to divorce or widowhood, was associated with cognitive decline. Using mixed-effects regression we found that the rate of cognitive decline increased after relationship termination (widowhood: □ = -0.587, p <0.001; divorce: □ = -0.221, p <0.001), supporting mechanism (5). Using HRS data for respondents and their spouses’ mental and physical health, health insurance, and activities of daily living, we also find support for mechanisms (1) and (3). Relationship termination is a critical juncture in a person’s life course that has multiple implications and may, ultimately, worsen patients’ conditions.

Current and Future Directions in Minority Aging: Embracing Interdisciplinary Models

Racial/ethnic disparities in health among older adults are well-documented. More research is needed to clarify the complex and multifactorial mechanisms underlying these associations. This symposium will feature research that employs innovative theoretical and methodological approaches to understand the biopsychosocial mechanisms that underlie racial/ethnic disparities in older adults’ health and determine sources of within-group heterogeneity in minority aging. Dr. Forrester will integrate stress biology and intersectionality to demonstrate the importance of stress and resilience (e.g., John Henryism) with biological aging within Black adults participating in the Coronary Artery Risk Development in Young Adults (CARDIA) Study. Dr. Brown Hughes will present innovative research using data from the African American United Memory and Aging Project (AA-UMAP) on the importance of Alzheimer’s disease-specific knowledge and perceptions among Black older adults. Dr. Gamaldo will employ a within-race approach to understand how knowledge and perceptions of Alzheimer’s disease and related dementias (ADRD) shape cognitive performance among Black older adults in the AA-UMAP study. Dr. Mitchell will use Health and Retirement Study data to explore the role of midlife stress exposure in accounting for racial disparities in trajectories of cognitive functioning. Drs. Thierry and Farmer will use HRS data to examine how psychosocial resilience (e.g., mastery) affects the relationship between perceived neighborhood conditions (e.g., disorder) and cognition among Black older adults. This work highlights the importance of applying an interdisciplinary lens to move the study of minority aging forward and ultimately, to reduce the unnecessary burden of morbidity and mortality among minoritized groups.

The Interplay Between Stress Related Genes and Its Role in Human Longevity: Insights for Translational Studies

Human lifespan is a multifactorial trait resulted from complicated interplay among many genetic and environmental factors. Despite substantial progress in clarifying many aspects of lifespan’ variability the mechanism of its multifactorial regulation remains unclear. In this paper we investigate the role of genes from integrated stress response (ISR) pathway in such regulation. Experimental studies showed that persistent cellular stress may result in cellular senescence (for proliferating cells), or in apoptosis (for post-mitotic cells) which may affect health and lifespan in laboratory animals. These studies also showed which ISR genes are likely to interplay to produce joint effects on these traits. Note that in humans, the interplay between these genes does not necessarily influence these traits. This is because biological mechanisms regulating these traits in laboratory animals and humans may differ. This means that, when possible, the experimentally detected connections promising for human applications, should be verified using available human data before their testing in expensive clinical trials. In this paper we used HRS data to test connection between SNPs from the EIF2AK4 gene that senses cellular stress signals and the DDIT3 gene from the apoptosis regulation part of the ISR. We found genome wide significant associations between interacting SNPs from these genes and longevity. This result shows that available human data may be successfully used for making important steps in translation of experimental research findings towards their application in humans. Following this strategy may increase efficiency of clinical trials aiming to find appropriate medications to promote human health and longevity.

Stress Mediating Genes in Aging, Health, and Longevity Traits: Effects of Multiple Interactions

Persistent stress of external or internal origin accelerates aging, increases risk of aging related health disorders, and shortens lifespan. Stressors activate stress response genes, and their products collectively influence traits. The variability of stressors and responses to them contribute to trait heterogeneity, which may cause the failure of clinical trials for drug candidates. The objectives of this paper are: to address the heterogeneity issue; to evaluate collective interaction effects of genetic factors on Alzheimer’s disease (AD) and longevity using HRS data; to identify differences and similarities in patterns of genetic interactions within two genders; and to compare AD related genetic interaction patterns in HRS and LOADFS data. To reach these objectives we: selected candidate genes from stress related pathways affecting AD/longevity; implemented logistic regression model with interaction term to evaluate effects of SNP-pairs on these traits for males and females; constructed the novel interaction polygenic risk scores for SNPs, which showed strong interaction potential, and evaluated effects of these scores on AD/longevity; and compared patterns of genetic interactions within the two genders and within two datasets. We found there were many genes involved in highly significant interactions that were the same and that were different within the two genders. The effects of interaction polygenic risk scores on AD were strong and highly statistically significant. These conclusions were confirmed in analyses of interaction effects on longevity trait using HRS data. Comparison of HRS to LOADFS data showed that many genes had strong interaction effects on AD in both data sets.

Functional limitation in later-life: the impact of sips, socialization, and sadness

Objectives Findings on the health impact of moderate alcohol use in later-life have been discrepant. Unaddressed mechanisms of action and validity concerns may contribute to contrasting results. Several studies have demonstrated a beneficial relationship between moderate alcohol use and functional ability. Research is needed to identify mechanisms of action that explain this relationship, while employing methods to reduce common validity concerns in alcohol use research. This study examines social interaction and depressive symptoms as serial mediators in the relationship between moderate alcohol use and functional limitation in later-life. Method HRS data from 2012–2016 were utilized (n = 1,474); heavy drinkers, adults younger than 65, and respondents with inconsistent alcohol use from 2008 to 2016 were excluded. Hypotheses were evaluated using a longitudinal serial mediation model with bias-corrected bootstrapping. Results Results indicated that, in the context of demographic variables, medical burden, and previous functional limitation, the relationship between 2012 moderate alcohol use and 2016 reduced functional limitation is only present when considering the mediating effect of 2014 depressive symptoms as a single mediator, and 2012 social interaction and 2014 depressive symptoms as serial mediators, respectively (total effect model R 2=.167). Conclusions Data indicate that relationships between moderate drinking and reduced functional limitation are explained through increased social interaction and subsequent reduced depressive symptoms. Results identify social interaction as an accessible treatment target to address depressive symptoms and functional limitation, and support increased assessment of IADLs in adults experiencing depressive symptoms (and vice versa) to facilitate early treatment/prevention of functional and emotional decline.

NUMBER OF FRIENDS AND THE RISK OF DEVELOPING DIABETES: A REPLICATION USING NSHAP AND HRS

The epidemic increase in diabetes prevalence (primarily type 2) is a public health crisis. We hypothesized that the rates of movement among diabetic states depend in part on one’s social relationships and environment. Using population-based samples from both NSHAP and HRS, collected in 2005–15, we found that having more friends was associated with a lower risk for acquiring diabetes over the next 4-5 years. As an independent replication, separate logistic models for NSHAP and HRS data yielded similar odds-ratios for the protective effect of having friends (OR = 0.82 and 0.92 respectively), adjusting for gender, age, race/ethnicity, and BMI. This effect was concentrated entirely between 0–4 friends; differences in the number of friends above 4 were not associated with differences in diabetes risk.

Systematic Mortality Improvement Trends and Mortality Heterogeneity: Insights from Individual-Level HRS Data

Providers of life annuities and pensions need to consider both systematic mortality improvement trends and mortality heterogeneity. Although how mortality improvement varies with age and gender at the population level is well studied, how trends vary with risk factors remains relatively unexplored. This article assesses how systematic mortality improvement trends vary with individual risk characteristics using individual-level longitudinal data from the U.S. Health and Retirement Study between 1994 and 2009. Initially a Lee-Carter model is used to assess mortality improvement trends by grouping individuals with similar risk characteristics of gender, education, and race. We then fit a longitudinal mortality model to individual-level data allowing for heterogeneity and time trends in individual-level risk factors. Our results show how survey data can provide valuable insights into both mortality heterogeneity and improvement trends more effectively than commonly used aggregate models. We show how mortality improvement differs across individuals with different risk factors. Significantly, at an individual level, mortality improvement trends have been driven by changes in health history such as high blood pressure, cancer, and heart problems rather than risk factors such as education, marital status, body mass index, and smoker status.

Victim Characteristics of Investment Fraud

Investment fraud constitutes a major problem in the United States. While several studies have investigated various aspects of fraud, none have analyzed victim characteristics of investment fraud. This study posits five fraud languages that, when used by fraudsters, shut down the perceived need to conduct due diligence in their victims: perceived success; air of familiarity; claim to authority; noble pursuits; and framed authenticity. Using the Health and Retirement Study data from 2008, 2010, and 2012, the authors found that respondents who were male, were better educated, were single, younger, and possessed greater net worth disproportionately reported being defrauded in the past five years. Additionally, deploying principal component analysis (PCA) revealed that respondents who made responsible spending decisions, specifically those who were more likely to spend wealth over their entire lives as opposed to spending greater percentages of wealth on short-term needs, were less likely to report fraud. This study lays the groundwork for linking the five fraud languages to various factors such as financial literacy and dependency arising from questions within the HRS data set. Implications for such findings include protecting individuals near retirement from fraud and spreading public awareness about the importance of due diligence in the investment decision process.

Determinants of Hearing Aid Use Among Older Americans With Hearing Loss.

Hearing loss (HL) is common among older adults and is associated with significant psychosocial, cognitive, and physical sequelae. Hearing aids (HA) can help, but not all individuals with HL use them. This study examines how social determinants may impact HA use. We conducted an explanatory sequential mixed methods study involving a secondary analysis of a nationally representative data set, the Health and Retirement Study (HRS; n = 35,572). This was followed up with 1:1 qualitative interviews (n = 21) with community participants to clarify our findings. Both samples included individuals aged 55 and older with a self-reported HL, with or without HA. The main outcome measure was the proportion of participants with a self-reported HL who use HA. Analysis of HRS data indicated that younger, nonwhite, non-Hispanic, lower income, and less-educated individuals were significantly less likely to use HA than their referent groups (all p values < .001). Area of residence (e.g., urban) were not significantly associated with HA use. Qualitative findings revealed barriers to HA included cost, stigma, vanity, and a general low priority placed on addressing HL by health care providers. Facilitators to obtaining and using HA included family/friend support, knowledge, and adequate insurance coverage for HA. Many socioeconomic factors hinder individuals' ability to obtain and use HA, but these obstacles appeared to be mitigated in part when insurance plans provided adequate HA coverage, or when their family/friends provided encouragement to use HA.

Closing Down the Shop: Optimal Health and Wealth Dynamics Near the End of Life

Near the end of life, health declines, mortality risk increases and curative is replaced by uninsured long-term care, accelerating the fall in wealth. Whereas standard explanations emphasize inevitable aging processes, we propose a com- plementary closing down the shop justification where agents’ decisions affect their health and the timing of death. Despite preferring to live, individuals optimally deplete their health and wealth towards levels associated with high death risk and indifference between life and death. Reinstating exogenous aging processes reinforces the relevance of closing down. Using HRS data for elders, a structural estimation of the closed-form decisions identifies and tests conditions for these strategies to be optimal and confirm their economic relevance. We also discuss why policy intervention to reduce the incidence of closing down, although feasible, is not warranted.

The Accuracy of Economic Measurement in the Health and Retirement Study.

We assess the quality of the HRS's measures of out-of-pocket medical spending and health insurance premia, both in the "core interviews" and in the "exit interview" data. We provide detailed evidence on the quality of the HRS insurance premia data, and we compare the HRS exit data to exit data in the MCBS. We document how changes in survey questions, including the introduction of "unfolding brackets," affect the HRS measures. We document what we believe are errors in the HRS imputations and provide some suggestions for improving the accuracy of some imputed variables. Overall, we find the HRS data to be of high quality. However, we believe that many interesting variables in the HRS are under-utilized because users must perform imputations themselves.

Health Trajectories of Older Americans and Medical Expenses: Evidence from the Health and Retirement Study Data Over the 18 Year Period

This study investigates the long-term relationship between individuals’ health state changes over time and burdens due to out-of-pocket medical expenses (OOP) in later years. We kept track of 5540 individuals’ health trajectories and their accumulated OOP using the HRS data from 1992 to 2010. American adults between 50 and 70 years old spend on average $27,000 on OOP, and have five common health trajectory patterns (Multi-Morbidity, Co-Morbidity, Mild Disease, Late Event, and No Disease). However, their OOPs differed substantially depending on the pattern of health trajectory. The most costly pattern of Multi-Morbidity needed $18,823 more than the least costly No Disease pattern. Older adults with the most costly pattern spent most of OOP on either prescription drugs or doctor/dental visits. Additionally, we found that the OOP burden of prescription medications was substantially relieved by the Medicare Part D implementation. These findings have several important implications for individuals, financial educators, and policy makers.

EFFECT OF OBESITY ON ACTIVE LIFE EXPECTANCY IN THE UNITED STATES: CHANGES OVER TIME

In recent decades, life expectancy increase has slowed in the United States. Also, there have been striking increases in the prevalence of obesity, decreases in the age of obesity onset, and increases in the length of time people are living with obesity. In addition, the link between obesity and mortality appears to have weakened in more recent years. These trends indicate value in revisiting the links between obesity and active life expectancy. Building on earlier work using HRS data from 1993 to 1998 for the 70 plus population in the United States, we use HRS data from 2004 through 2008 to examine how the links between obesity, mortality, disability and subsequent active life expectancy have changed. We find that in the later cohort obesity is less related to mortality, more to disability, and remains a strong determinant of active life.

Personality Traits and Financial Decisions of the Households

Abstract. Who makes the financial decisions in the household and how does the identity of the decision-maker affect the subsequent financial choices? In this paper I show that it is not necessarily the most able or resourceful, but the most determined and organized of the spouses takes the responsibility for financial decisions. I also observe an age-related shift in financial management of the household finances: for older households, higher score on conscientiousness is one of the most significant predictors of stock ownership, as well as of the higher share of wealth invested in stocks, while for younger ones explanatory power lays with cognitive abilities and education. Also, the influence of the spouse on the financial decisions with respect to stock market is documented. The study uses HRS data to directly identify the decision-maker in the household and considers the effects of personality traits above and beyond cognitive abilities, risk aversion, demographic and economic factors. On one hand, this paper contributes to the discussion regarding personality traits and their explanatory power with respect to financial choices of households. On the other hand, it adds to the well-established literature on the intra-household decision-making, considering the personality traits of both spouses as the determinants of the decision-making power distribution within the family.

Does Social Security Continue to Favor Couples?

While dramatic increases in women’s labor supply and earnings have led to a substantial decline in the fraction of women eligible for spouse benefits at retirement, most wives still receive a survivor benefit, as wives still typically have lower earnings than their husbands and live longer. Using the MINT microsimulation model and the HRS data linked with Social Security administrative earnings records, this paper examines the extent to which Social Security continues to favor couples and will do so in the future.The paper found that: While the Old-Age and Survivors Insurance program still distributes lifetime income from singles to couples, the transfers appear to be shrinking over time. Nevertheless, couples are still projected to have a higher benefit/tax ratio, a lower median net tax rate, and a greater likelihood of receiving positive net transfers from the system compared to those who are never married or divorced. The increased labor force participation and earnings of women have contributed significantly to the decline in redistribution from men to women, and from singles to couples, while the effect of declining marriage rates has only a modest effect. The policy implications of the findings are: Family benefit provisions within the Social Security program can have a significant impact on various measures of redistribution. Policymakers may find the results of this paper helpful in evaluating any reform proposals that would change these provisions.

Motives for Bequests within the Middle Class

The life-cycle model of household behavior forms the basis for most economic analysis of Social Security, private pensions, and retirement. This project seeks to extend the usefulness of the life-cycle model by considering the role of middle-class inheritances and bequests. We use HRS data. Prior work by the authors identifies key information in the HRS on the sources of private intergenerational transfers, and it shows that the frequency of couples’ inheritances from both spouses’ family lines is higher than random behavior would imply. Using additional HRS data on the ratio of parent-to-child lifetime incomes, we analyze the motives behind HRS bequests. We find support for an unintentional transfer model in which bequests arise from residual, unspent parent life-cycle resources. And, we show that our model can account for the frequency of dual inheritances that earlier work revealed.