___________________

¹ This chapter provides a summary and integration of the findings from the preceding chapters. Where relevant, we reference the report chapter that contains a more detailed discussion of the evidence underlying the findings and conclusions presented. Where the findings of specific studies are discussed, the original references are provided.

The first conclusion points to the need for future research to recognize the ways in which age bias and social and economic inequality, as well as changes in them over time, set the contexts in which individual preferences, expectations, and constraints on work and retirement behaviors at older ages are shaped. Individuals’ preferences and expectations for work as well as their ability to realize these preferences are constrained by the opportunity structures in which they live. These structures do not suddenly emerge at older ages but are the result of social and economic inequalities that shape economic and health outcomes throughout the life course and as individuals move through changing historical contexts. As a result, adults with higher levels of education and wealth experience fewer constraints that prevent them from extending their working life beyond traditional retirement age, and education-based gaps in employment rates increase with age among older adults (Chapters 2 and 4).

These differences in employment rates have implications for what we know about work at older ages. Because older workers who remain in the labor force are on average more affluent than older adults in the population, data and research that focuses on older workers will reflect the relative affluence of this population. However, these findings may not represent the experiences of less affluent populations that face a more restrictive opportunity structure. The second conclusion draws attention to the need to better understand the heterogeneity within the older worker population to ensure that the experiences of vulnerable populations are well-represented in research and policy discussions. Considering the diversity of experiences and outcomes within the older worker population, as well as the ways opportunities are shaped by inequality throughout the life course, can provide important research insights—even when research is not explicitly focused on social and economic inequality.

As an example, consider the efficacy of the Age Discrimination in Employment Act (ADEA) and other state age discrimination policies reviewed in Chapter 8. Though on balance the evidence suggests that—outside of periods of significant economic recession and disruption, such as the Great Recession—strong age discrimination laws are effective in improving employment among older workers, the limited number of studies that have examined whether their effects differ by gender or race suggest that they have little or no effect on reducing discrimination among older women or racial minorities. Moreover, when age discrimination laws are weak (i.e., place more restrictions on the damages plaintiffs can seek), most plaintiffs under the ADEA are highly paid White male middle-managers. These differences are especially concerning given the higher rates of hiring discrimination experienced by older women and Black adults (Chapter 6). They suggest that more affluent White male workers may experience greater protections from age discrimination than other older workers, compounding the effects of labor market advantages they experience throughout their working lives. However, the reasons underlying these differences are not yet well understood.

Taken together, these two overarching conclusions provide a framework though which the research agenda laid out in this chapter can be viewed. They can serve as a call to ensure that future research considers the heterogeneity of experiences within the older worker population, including the diverse ways in which work and retirement outcomes are shaped by broader contexts of age bias and social and economic inequality throughout the life course. Considering this heterogeneity will provide a more comprehensive and robust understanding of the benefits and disadvantages of extending work lives, which in turn can inform the development of policies that enable work at older ages in ways that account for the specific challenges vulnerable populations face and, therefore, improve their well-being.

DEFINING A RESEARCH AGENDA

These overarching conclusions provide the context for a future research and data collection agenda. Work and retirement preferences, expectations, and behaviors are shaped by many factors that operate at different levels of analysis, including continuity and change in individual and family characteristics and resources, workplace policies and practices, age discrimination, labor market opportunities, and social policies. When considered as a whole, the

findings within the preceding chapters reveal a research agenda that, if enacted, would provide a deeper and more comprehensive understanding of the employment experiences of older workers and the constraints that shape their opportunities for work at older ages. They identify three lacunae in existing scholarship that cut across multiple levels of analysis and are the most promising areas for future research:

1. the employer–older employee relationship;
2. work and resource inequalities in later adulthood; and
3. the work-health-caregiving interface.

In the remainder of this chapter, each of these topics is considered in turn, drawing on the findings of the previous chapters to identify key gaps in the literature and research questions that remain unaddressed. It is important to note that many of these research gaps remain not because of a failure of imagination on the part of researchers in this area, but because their efforts are constrained by the limitations of the tools available to them. Thus, this chapter highlights research areas where the development of methodological innovations and improvements to current data infrastructure may be required.

The Relationship Between Employers and Older Employees

The role workplaces play in affecting older adults’ work and retirement pathways has received relatively less research attention than the role individuals play (Chapter 5). Organizations are the crucial link between macro-level public policies and individual-level outcomes among workers. The employer-employee relationship translates national policies into organizational practices that set the stage for individual decisions to continue in their current job, accept a similar position at another firm, transition to bridge employment or part-time work, or leave the labor force. To ignore this relationship in understanding late-career behavior and outcomes has major repercussions for getting a clearer picture of the way older adults approach workforce participation and retirement.

Employer discretion is particularly important in determining what types of workers are hired, trained, and retained, as well as the types of workplace policies and programs that are put in place to recruit and retain workers within labor markets. Employer decisions directly affect the extent to which workers experience longer, healthier, and more engaged working lives. Moreover, the preretirement performance and well-being of older workers cannot be understood without sufficient knowledge of workplace practices and policies. Understanding the role of employers and employees in facilitating longer working lives is paramount today, in light of the unprecedented aging of the population and the resulting pension reforms that are being implemented to deal with this demographic reality. Important areas for future research include the implementation of workplace policies; policies and practices that affect work and retirement; and the role of age discrimination.

Implementation of Workplace Policies and Practices

The employer-employee relationship in the workplace is shaped by the interests of both the employer and the employee. These interests can differ or align depending on the workplace issue, for example whether it is safety procedures or compensation. The institutional environment also affects the interests of employers and employees. Laws that distribute rights and power to employers or employees influence the ability of either party to enact practices that reflect their interests, as does the quality of labor markets, which can enable or inhibit workers from finding new positions and employers from hiring new workers. Crucial to understanding the purpose of various workplace policies and practices affecting older workers is the employer-employee relationship and its institutional

environment. Therefore, we encourage researchers to explicitly account for employer and employee interests and the institutional context when analyzing workplace policies and practices affecting older workers.

Employee voice is one key aspect of the employer-employee relationship that can shape workplace policies and practices (Chapter 5). The extent to which employees have a say in their work can influence the type of age-related or age-neutral practices employers implement. For example, employees may prefer particular flexible scheduling arrangements or certain ergonomic practices. Employee voice can be part of independent collective representation that engages in negotiations with employers, or it can occur when individual employees are able to express their concerns and opinions directly to employers. Greater employee voice at the workplace can be instrumental in expanding work options for older employees, such as part-time work, remote work, or partial retirement. Moreover, greater employee voice can help reduce inequalities by fostering greater access to practices across ages and historically disadvantaged subgroups.

Forms of employee voice often vary by country, reflecting different laws, norms, and levels of union activity. Thus, international comparative studies can be particularly useful in demonstrating the influence of employee voice on the types of practices implemented within organizations. We encourage researchers to examine the role that employee voice can play in reducing inequality in access to and use of various workplace practices that affect older workers.

When employers introduce new workplace practices that benefit older workers, they can do so in either age-specific or age-neutral ways. Unfortunately, only limited research compares the effectiveness of age-specific and age-neutral practices on the employment decisions of older workers (Chapter 5). Age-specific practices are tailored to older workers and can be useful in meeting their specific needs, but these practices run the risk of stigmatizing older workers as a “special group” and encouraging discrimination against older employees (Chapters 5 and 6). Moreover, these policies can be problematic to the extent that they reinforce stereotypes or raise legal concerns of promoting differential treatment based on age, making the employer vulnerable to claims of age discrimination (Chapter 8). Age-neutral practices are designed for all workers and therefore avoid age-based stigmatization; however, when the needs of older workers differ from those of younger workers, such practices risk losing the effectiveness they would have had had they been specifically designed for older workers. More research and organizational data sources are needed to better understand the implications of these tradeoffs.

Policies and Practices That Affect Work and Retirement

Older employees are more likely to report a desire for flexible work arrangements than younger workers (Maestas et al., 2018), and that greater flexibility would increase their ability to remain in the labor force at older ages (Anderson et al., 2020). Flexibility can also help organizations adapt their operations, but differences in how, why, and for whom these arrangements are available can lead to very different employee experiences (Chapter 5). Methods of introducing flexibility into work practices, such as through remote work, flexible hours, or alternative schedules, can be voluntary, chosen by workers; involuntary, implemented by managers or employers for business reasons; or mandated by the government due, for example, to a pandemic or natural disaster. These differences can affect whether the introduction of flexible arrangements improves the retention of older workers and the working conditions of workers more generally. Some forms of flexibility can lead to increases in work hours or reduced

employee control over work schedules, which in turn can lead to reduced retention of older workers. More research is needed on the effects that voluntary and involuntary flexibility in hours, in the ability to take time off, and in work location have on older workers’ decisions to exit or remain employed.

Clearly identifying both the type of flexibility and whether it is voluntary will be crucial for ascertaining the effects of flexible work arrangements on the retention of older workers. Only limited research examines the effects of different forms of flexible arrangements on the timing of retirement. Europe offers some comparative models for allowing reduced hours of work as a form of partial retirement, a practice more commonly encountered there than in the United States (Chapters 5 and 8). Moreover, little is known about how experiences of remote working during the COVID-19 pandemic serve to increase older workers’ preferences for continued work participation, including their demands for partial retirement schemes. Research linking flexible work practices to later-life work could investigate not only exits from the workforce, but also the effects of these arrangements on older workers’ health and well-being, work-life balance, and ability to provide caregiving (Chapter 4). These outcomes are important in their own right, but they are also important as possible mechanisms through which flexible work arrangements influence individual pathways into and out of the workforce.

Employers also influence the ways in which technological innovations are introduced in the workplace. Innovations in technology and automation are rapidly changing the way work is carried out in organizations today (Chapter 5). There is evidence that such innovations can support older workers in their efforts to continue working in a variety of ways. For example, technology can make jobs safer by monitoring repetitive tasks or providing lift and positioning technology on assembly lines or using “cobots” to aid workers’ strength (Coughlin, 2020). On the other hand, there is evidence that technological innovations are eliminating the need for some jobs, especially those involving repetitive tasks that do not require higher education (Chapter 7). The elimination of these occupations, in turn, reduces the number of employment opportunities for workers without a college degree and increases economic inequalities. Although it is difficult to study an ever-moving target (the technologies themselves are changing rapidly), it is imperative to better understand which technologies are beneficial and which are destructive to making continued work in later life a reality for those who need it or would prefer to continue working. Research examining the implementation of technology in the workplace, including the reasons for its introduction and the impact it has on the health, well-being, and employment opportunities of older workers, is important for understanding both its benefits and its role in limiting older workers’ job opportunities.

Worker training and skill plays an important role in employee retention at older ages. Keeping skills up to date is critical for older workers who are interested in working longer (Chapters 5, 7, and 8). These human capital endowments are important factors that employers use to make decisions regarding practices that can increase the retention of older workers or expand their roles within an organization. Yet employee access to training programs depends on many factors, including occupation, manager discretion, and individual incentives to participate (Chapter 5). Older workers are often excluded from training and development programs due to employer assumptions that older workers are uninterested in these programs and will not provide a sufficient return on their investment in training (Chapters 5 and 7). We do not yet fully understand the distribution of training and development access across employees within organizations or the consequences of access to these programs in shaping the timing of retirement.

various less-examined ways. If the biases are ambivalent, implicit, ambiguous, and latent, then people cannot individually track their own biases in hiring, managing, or promotion decisions; instead, a firm should watch for patterns that might result from many instances of unexamined bias. Unexamined ageism needs to be explored with more direct evidence, especially at work. Because undertaking such measurements is not straightforward, it requires development.

Implicit attitudes toward older people are, on average, negative, but only moderately so, and they have remained stable over time (Chapter 6). At the same time, the social acceptability of expressing explicit ageism appears to have declined over time. Ageist attitudes’ slight negativity masks a mixed stereotype: expecting an older person to be incompetent but well-meaning. However, age can interact with other individual characteristics (e.g., gender, race, or sexuality) to produce stereotypes of elder subtypes that differ from the “incompetent but well-meaning” stereotype. Documenting unexamined workplace biases might include identifying whether, when, and how:

1. an original, unfiltered attitude or a newer, socially desirable, revised attitude guides age-related workplace behavior, including discrimination;
2. implicit ageism results in unmonitored behavior that unintentionally leaks a negative attitude;
3. mixed ageist stereotypes (“doddering but dear”) predict ambivalent (e.g., patronizing) behavior; and
4. elder subtypes (e.g., “grandmother,” “recluse,” “statesman”) might increase the accuracy of predicting biases and their effects.

Additional research is needed to assess whether each subtle form of bias might require different forms of detection and mitigation.

Despite evidence of biased attitudes about older workers (Chapter 6) and age-based discrimination in the workplace (Chapters 6 and 8), the causal chain that produces age discrimination is incomplete, because while observers can report their stereotypes and prejudices, only the targets report discrimination (Chapter 6). The current evidence base is missing the link between supervisors’ and co-workers’ reported age-related preferences (attitudes, prejudices) and older workers’ reports of discrimination: coworkers’ and managers’ reported and observed discriminatory behavior toward older people. There is compelling evidence from field experiments that older workers suffer from age discrimination in hiring, demonstrating that stereotypes about older workers do affect employment decisions, which constrains labor demand for older workers. Discrimination against older workers seems to be reinforced by negative stereotypes about them. However, there is not the same kind of rigorous evidence on age discrimination along dimensions such as promotion, selection for training, and terminations (Chapters 6 and 8). Researchers thus far have not developed methods that generate rigorous evidence on age discrimination along dimensions other than hiring; this is a critical challenge. Funding agencies and foundations could encourage researchers to think outside the box about how to bring rigorous—likely experimental—evidence to bear on these other dimensions, including, for example, running experiments inside organizations.

Within workplaces, age-related biases operate not only within interactions at the interpersonal level but can infuse the broader organizational climate (Chapters 5 and 6). Consistent with the evidence and conclusions regarding ageism, research is needed to understand inclusive organizational policies that mitigate against age-related biases entering into decisions at all levels—hiring, supervising, and promoting. Moreover, organizational climate determines whether older workers feel valued. An inclusive culture makes older workers (and everyone else) feel that they belong. If latent prejudice cannot be eliminated entirely, organizations can still develop policies that help to keep it from creating an unfair and unwelcoming work environment.

Organizations often describe themselves as having a culture of inclusion, but investigations of organizational culture necessitate better conceptual and operational definitions of inclusive organizational policies, as well

as multi-level models of their antecedents, indicators, and consequences (Chapters 6). These factors link to other efforts for workplace flexibility, voice, and retention (see discussion above). In addition to developing methods of measuring organizational culture, research should assess whether interventions implementing policies to improve the inclusivity of workplace cultures actually create a more inclusive culture for older workers. Findings to date are mixed, and they neglect the question of whether age-discrimination regulations affect the existence and experience of age-based discrimination in the workplace; a later section returns to this point.

Organizational decisions affect the assessment, formal or informal, of the productivity of older workers, and there is also concern that age bias in evaluations of performance and productivity negatively affect older workers. Productivity is hard to measure, causally ambiguous, sensitive to level of aggregation, and distinct by job type, but on balance research does not show clear age-related productivity declines (Chapter 6). Older workers generally underperform relative to younger workers on speed (fluid intelligence) but outperform their younger counterparts on certain other performance-related measures, such as experience (crystalized intelligence) and workplace citizenship (social intelligence), yet researcher and employer evaluations of employee performance often omit these latter factors because they are more difficult to measure. As a result, assessments of older workers’ contributions may be biased, because these performance strengths are seldom examined and often discounted by both employers and researchers. Research that evaluates the impact of the exclusion of such older- worker strengths on researcher, supervisor, and peer estimates of worker productivity and performance would add precision to studies of older workers’ productivity and workplace performance as well as the value of retaining older workers.

Just as complex are the interrelationships between measures of worker “age,” worker character, job factors, and performance, which all contribute to analyzing expectation accuracy regarding performance. For example, as discussed in Chapters 5 and 6, in practice chronological age confounds the effects of the number of years lived, the generation into which an individual is born, job tenure, and work experience. There is a real need for scholarship using data that can filter out age effects from the many other determinants of day-to-day productivity. Determining the independent effects of each of these conceptually distinct factors is difficult, and more research is needed to clarify these complex concepts. Similarly, the way age intersects with other identities also contributes to underestimating the contributions of older workers. For example, there is insufficient understanding of the potentially stronger age discrimination experienced by older women compared to older men (Chapters 6 and 8). Workers who face disadvantage due to their race, ethnicity, sexuality, or immigrant status might experience multiplier effects as they age—or conversely, become invisible and immune to age effects (Chapter 6). Research needs to consider age-related intersectionality in the workplace, seeking both patterns and distinctive effects on older-worker pathways.

Older workers’ reports of discrimination covary with job dissatisfaction and poor health (Chapter 6), though the evidence of causality awaits better designs because third variables, such as depression, might explain both. Research that includes such covariates (e.g., optimism, personality, support, intersectional identities) will likely

clarify and isolate the relationships among perceived workplace discrimination, health, and labor force participation. Research needs to establish the presence of a causal relationships between these factors.

Given the wide perception of age bias, mitigation efforts are needed at all levels of decision-making in organizations today (Chapter 6). To some extent, Equal Employment Opportunity Commission requirements capture biased patterns but not the variables that might explain them or show how to mitigate them. Organization-level research needs to address how and when organizations can monitor the emergence of bias most effectively. There is a need for research that examines inclusive organizational polices designed to mitigate against biases entering into decisions at all levels—hiring, supervising, promoting—and that addresses the extent to which such policies do in fact increase employees’ perception of the workplace as being fair and inclusive.

New Data Collection Strategies

Addressing these research gaps regarding workplace practices requires gathering data on actors in the employment relationship, as well as understanding the organizational context and the perspectives of managers and older workers within that relationship. The importance of the organizational context on employee responses to and outcomes of any policy and/or practice underscores the necessity of having matched employer and employee data over time. Most of the existing panel surveys sponsored by the U.S. federal government focus on sampling individuals or households; such surveys might inquire about participants’ work, but they examine little about the context of their work or even, in some cases, their industry sector.

To date, research on workplace policies and practices in the United States has typically relied on access granted by individual companies or establishments within industries. While this approach has resulted in insightful research, such findings may not be generalizable beyond the organizations studied. More importantly, such access is very difficult to obtain in the United States and even more difficult to sustain over time. Negotiating individually with companies for access to survey managers and workers is time-consuming and requires demonstrating the benefits of the research to a specific company rather than the more general contribution to the advancement of knowledge. In addition, certain topics, such as discrimination, are legally sensitive, making it difficult to gain company consent to the participation of employees and managers.

Government surveys have a much greater chance of obtaining consistent data from establishments and workers over time. However, federal surveys by the U.S. Census Bureau, such as the Longitudinal Employer-Household Dynamics survey and the Management and Organizational Practices Survey, are designed with very limited organizational questions. Neither survey asks about human resource practices or age-related workplace practices, so neither addresses the fundamental question of access to and use of various practices at the workplace. Understanding the ability of individuals to access and use practices is critical to analyzing practice impact on older workers.

The National Compensation Survey by the Bureau of Labor Statistics suffers from the same problem. Although it includes a question about flexible schedules, it does not go beyond this practice, nor does it address access to or use of the practice. A more recent data set, called the Census-Enhanced Health and Retirement Study, links a range of establishment characteristics to the Health and Retirement Survey, focusing on older workers. This approach has some promise, but the establishment questions do not address key age-related workplace practices, focusing more on the economic characteristics of establishments.

In sum, no nationally representative longitudinal panel exists that samples U.S. workplaces and also contains multilevel matched data between employers and workers. To study the multi-layered and dynamic impact of workplace practices on work and retirement pathways, it would be invaluable to establish a panel survey or modifying an existing federal survey with sufficient variation at both the workplace and individual levels. Such a panel survey should collect data from a random selection of workplaces from the population of employers, covering all sectors and all sizes of workplaces. Within each workplace, data should also be collected from a random sample of older workers to directly gauge how older workers react to various workplace practices over time.

also more likely to work at older ages (Chapter 2), providing an additional source of income that increases their economic advantage relative to those with lower socioeconomic status. However, there is currently little research examining the ways in which disparities in employment patterns contribute to social and economic inequality between older adults.

Answers are not likely to be simple. Multiple identities, defined by characteristics beyond age, intersect to create distinct opportunities and constraints, shaped by forms of discrimination that play out at all life stages and accumulate in later-adult-workforce advantages and disadvantages. Disparities across demographic groups in work-related resources and options in later adulthood are found in health, caregiving responsibilities, household structure, financial literacy, job experience and skills, job insecurity, employment opportunities, lifetime earnings, wealth, and access to employer-provided pensions and health insurance (Chapter 4). The next sections, respectively, suggest research needed to address disparities from three perspectives: life-course dynamics that produce inequality throughout the lifespan, inequality in access to employment opportunities and experiences, and inequality in financial stability.

A Life Course Perspective of Inequality

Though research on older workers often focuses on workers ages 50 and over, the processes that structure unequal work and retirement pathways at older ages do not begin at age 50. Applying a life-course lens to work and retirement pathways means recognizing that the resources and work/retirement options available in mid- and later-life accumulate and are shaped through prior life-course experiences of advantages and disadvantages; the relationships in which individuals are embedded; and the historical, policy, organizational, and social contexts in which they occur (Chapters 3 and 4). A fundamental life-course theme is that transitions, such as the nature and timing of later adult exits from or reentries into paid work, occur within trajectories of experience that give them shape and meaning.

Adopting a life-course conceptual lens can promote a better understanding of disparities in later adulthood by improving our understanding of the processes that lead to inequality in retirement (Chapters 3 and 4). Individuals from different backgrounds come to later adulthood with divergent histories. People reaching conventional retirement ages with higher amounts of education, those in professional jobs, and those following orderly career paths are better positioned to follow their goals and motivations. Unfortunately, those unable to remain continuously in the workforce or in full-time, relatively secure jobs confront limited options. By shifting the research focus to how early experiences and contexts influence the current preferences, expectations, and constraints on opportunities for work of older workers, a life course perspective can identify how historical structural inequalities continue to affect inequality in work outcomes at older ages.

A second life-course theme concerns the interdependence of lives—not only across different life spheres but also across social ties. Couples often make decisions around retirement timing together, or in light of each partner’s preferences (Chapters 3 and 4). And family situational exigencies, such as the need to care for a grandchild or an aging family member, may well alter work and retirement preferences and expectations. A better understanding of these interdependent relationships would provide a more complete understanding of the contexts and constraints under which work transitions occur.

Yet a third life-course theme locates individual transitions and trajectories within historical time as well as within existing policy and institutional arrangements. Contextual embeddedness implies that life transition and

developmental trajectories occur under specific circumstances, including historical, policy, organizational, and social contexts. Individuals are embedded within multiple contexts (e.g., households/families, neighborhoods, states, nations), and individuals shape and are shaped by these contexts (Chapters 5, 6, 7, and 8). Failure to consider these contexts may lead scientists to put too much emphasis on the effects of individual characteristics and to fail to recognize the ways in which the effects of individual characteristics are constrained by the contexts in which they are embedded and the changes in these contexts over time.

Inequality in Work Opportunities

Though work at older ages can substantially improve the economic security of older adults, work opportunities are not always available. Opportunities to remain in the labor force often reflect the underlying economic inequality in opportunities by socioeconomic status and geography (Chapters 4 and 7). Labor market factors such as globalization and automation target specific occupations and locations, affecting work opportunities within some geographic areas or for some classes of workers, often defined by skill level or type, but also defined by gender and race-ethnicity due to labor market segregation (Chapter 7). These changes often reinforce existing economic inequalities, because displaced older workers are less likely to be reemployed.

Though there has been extensive research on the labor market effects of globalization and automation, this research has not specifically examined the impact of these changes on older workers (Chapter 7). Globalization leads to changes in the industrial and occupational structure of employment. The lower geographic and occupational mobility of older workers can leave them at a disadvantage when labor demand for the jobs they have occupied declines. In addition, in the United States (and other advanced economies), globalization may increase the premium for learning new skills. Broadly, countries with older workforces tend to be more likely to lose out, perhaps because of less skill adaptability within their labor force, but there is little research that has examined the direct impact this has on older workers. Similarly, the effects of automation on labor demand for older workers specifically, as well as the long-term implications of automation-related job loss at younger ages, are not known (Chapter 7). Some evidence indicates that certain types of automation, at least, increase demand for older workers, especially for those who adapt to the new technologies, but other types may have the opposite effect. Most research on the effects of automation is speculative at this stage, with only limited empirical evidence on recent changes and little research that goes beyond more “traditional” forms of automation to look at the automation of cognitive tasks by machine learning and artificial intelligence.

Globalization and automation have contributed to and expanded on the effects of growing regional disparities in economic outcomes in the United States, which have left rising numbers of people in areas with low employment rates, high unemployment rates, high rates of Social Security Disability Insurance receipt, and related social ills, which might have broader repercussions for their well-being (Chapters 7 and 8). The role of geographic mobility in reducing these disparities has weakened over time (Chapter 7). There are reasons to suspect that older workers, who are less mobile than their younger counterparts, are more likely to be affected by geographic disparities in these labor market changes. However, research has not focused on older individuals per se, leaving knowledge about these age-related disparities as purely speculative.

In general, more research is needed to understand the impact of globalization, automation, and geographic disparities on the work opportunities available to older workers, and how public policy might respond to mitigate adverse impacts. Moreover, little is known about the long-term impact of job loss at younger ages from these forces and how that could affect work at older ages by altering employment prospects, health, or take-up of Disability Insurance benefits (Chapters 4, 7, and 8). Research is needed on regional place-based policies that might boost employment in poorly performing areas and how such policies might begin to reverse the decline in employment, particularly among lower-skilled older men (Chapter 7).

Mismatch between workers’ skills and the skills employers demand can arise from economic dislocations and changes in the demand for skills. The effects of this mismatch can be worse for older workers, in part because of weaker incentives to invest in training or in moving (Chapter 7). Opportunities for retraining are often available through employers, but older workers are less likely to participate in these programs (Chapters 5 and 7). Organizational policies and practices occur in the context of larger public policies that may also have broader effects (Chapter 8), including for those who are not currently employed. The long-term effects of automation- and globalization-related dislocation could perhaps spur governments to invest more in the retraining of older workers, but the effectiveness of programs (using methods such as cost/benefit analyses) that help or encourage older workers to invest in skills that match employer needs is not known (Chapters 7 and 8). Some government programs in the United States try to help displaced older workers find new jobs, but their effectiveness has not been evaluated. Some kinds of retraining—like community college education in the United States and firm training subsidies in Germany—boost earnings and employment, although the social returns may be relatively low if workers are relatively close to retirement (Chapter 7). Additionally, programs in other countries—such as Japan—merit evaluation.

To better understand work and resource inequalities in later adulthood, we need rich information about the most vulnerable populations in the labor market. These include less-educated, minority, women, self-employed, and “gig” and informal workers, as well as workers with disabilities, those in poor health, those who are noncitizen immigrants, and those living in rural areas. In particular, we need to understand their labor transitions in and out of work and in and out of the formal and informal sectors, because these are likely to be more complex than those of more advantaged workers and are not always captured using common measures of employment (Chapter 2). We also need to understand any simultaneous participation in formal and informal sector jobs, involuntary job separations that lead to labor force withdrawal before normal retirement age, barriers to work at older ages, sources of income, multiple occupations, sporadic jobs, access to health care and other labor benefits, and eligibility and take-up rates of supplemental income support programs for these low-income older adults (Chapters 4 and 8).

Vulnerable older adults may face unique labor market entry barriers. These barriers may relate to their education level, age, race/ethnicity, gender, health status, and access to urbanized areas (Chapters 2 and 4). Because there is significant heterogeneity within these populations and in their experiences, a full understanding of their diverse work pathways requires comparatively large samples for analysis. It also requires more detailed survey questionnaires that can capture the complexity of their work experiences, including multiple occupations, sporadic or gig jobs, and frequent transitions into and out of work (Chapter 2). Previous qualitative and quantitative work may be extended to develop further understanding of these populations.

Social networks and social capital may play an especially important role in identifying employment opportunities for older adults with more tenuous and less formal contact with the labor market (Chapter 4). These social connections inform and enable varied paths through work and retirement. Only limited research has addressed the role of key social mechanisms in obtaining employment in mid- and later life; even less empirical research focuses on older racial and ethnic minorities across socioeconomic status. Much of existing research is focused on youths and adults whose social networks (both strong and weak social ties, social support and social leverage,

and homogenous/heterogeneous networks) play significant roles in labor force participation and job-seeking (Chapter 4). These factors all help to understand the process of finding meaningful and satisfying work, yet we know little about how these social mechanisms function with advancing age, as social networks retract, especially when individuals are pressed to retire or if they live in communities with few employment opportunities. These social resources may benefit older job seekers up to only a certain point; their efficacy may depend on the socioeconomic characteristics of the social network and the levels of trust and reciprocity among members of the group.

National, state, and local organizations are institutions that can help build social capital among older adults (Chapters 4 and 8). For example, formal volunteering could expand and strengthen social ties, while also bolstering various dimensions of health (Morrow-Howell et al., 2001; Gonzales and Nowell, 2017). These social and health resources generated and maintained by volunteering might then help older adults who need or want to work to acquire employment in later life. The pathways between paid work, retirement, civic engagement, and returning to the labor market are not fully understood. For example, some large companies provide volunteer opportunities to pre-retirees to help smooth the transition between paid work and retirement. Volunteering offers pre-retirees a purposeful role as a potential substitute for their work identity. Another pathway is probably more common: individuals transitioning from employment to retirement, engaging in some leisure, and then starting to volunteer. In either pathway, some volunteers transition out of retirement and go back to work (Carr and Kail, 2013; Gonzales and Nowell, 2017). The many pathways between work, civic engagement, and retirement have not been empirically identified, nor do we know which path optimizes health and retirement security. Knowing and tracing these pathways, especially across a diversity of populations, could help clarify the role of employment and civic engagement policies and practices in later life.

Inequality in Financial Security

Financial security is one of the most important considerations workers face in decisions to retire or continue working longer. Inadequate retirement savings may constrain older individuals in deciding when to retire (Chapter 4). There is considerable debate regarding whether Americans have adequate savings for retirement (Chapter 8). Studies that use an “optimal saving model” to estimate whether individuals are saving optimally for retirement find that the vast majority of households reach this goal, while studies that use the “target replacement rate model” report widespread inadequacy.² No matter which point of view is taken, however, experts agree that workers today face growing challenges in saving adequately for retirement due to the increased life expectancy, decreases in Social Security benefits as the full retirement age rises, changes in employer-sponsored pensions, declining real interest rates, and other changes that have increased the need for higher ratios of wealth to income (Chapters 4 and 8).

While there is no clear consensus on whether the population as a whole is saving adequately for retirement, there is growing concern about the large wealth gaps by race, ethnicity, and education (Chapter 4). Financial security in old age is more tenuous for members of historically disadvantaged groups due to their lower wealth. Lower lifetime earnings and less access to employer-sponsored pensions impede retirement savings. Safety net programs make ineligible those individuals with assets above a certain level, reducing incentives for low-income earners to save or accumulate above that level. A lack of financial literacy and access to financial institutions, financial products and services, and opportunities for building savings and other assets across the life course can be additional barriers.

___________________

² Optimal savings approaches model consumption over the life course to develop household-specific optimum wealth targets and compare these to savings behaviors to determine savings adequacy (Scholz et al., 2006). Target replacement rate models compare projected household retirement income from assets, Social Security, and pensions to pre-retirement standard of living (Munnell et al., 2006). A more detailed discussion appears in Chapter 8.

duration of providing care to family members often conflicts with employment and often leads to part-time work and forced retirement (Chapters 3 and 4). The interrelationship between caregiving and labor force participation is dynamic and complex as they jointly play out in later life. Differences in the type and intensity of the caregiving needs of the recipient as well as the likely interaction effects between the health, social, and economic status of the informal caregiver and that of the care receiver add to the complexity of this relationship. It is likely that there are interaction effects as well between working conditions, employment policies, and practices shaping resources for caregiving (Chapters 4, 5, and 8). The fluctuation of caregiving demands affects the timing of retirement and other exits from the labor force (Chapter 4). These individual, dyadic, and employment dynamics are further shaped by state and federal policies, such as paid family and medical leave (Chapter 8). Research has not yet modeled these conjoint later-life-course dynamics, nor examined how they differ across diverse groups of caregivers and care receivers.

Informal caregiving responsibilities influence both preferences for and the actual timing of labor force exits (Chapter 4). Older workers may prefer working beyond conventional retirement ages, but they face constraints because of caregiving and family demands. Amid great diversity of caregiving and work arrangements, longitudinal studies in the United States and abroad suggest that older informal caregivers often reduce hours at work, feel forced into retirement, retire earlier, and do not return to work after retirement (Chapter 4). Women are more likely to leave the labor market permanently, more likely to work part-time due to informal care, and more likely to take time off work due to the demands of care, even controlling for gendered occupations and industries, job tenure, and union membership (Chapter 4). As the intensity of providing care increases, the likelihood of returning to work decreases.

However, in contrast to this consistent finding, a small body of research associates informal care with delaying retirement or going back to work after retirement (Chapter 4). The costs of caregiving can push older workers to remain in or reenter the workforce to provide financial support. Research can examine the tradeoffs between the demands of informal caregiving and those of employment and how these dynamics influence retirement pathways.

Caregiving demands may also create barriers for public policies that aim at facilitating employment for older adults, such as the Senior Community Employment Program (Chapter 8). Such community service and work-based training programs for older adults usually require that participants invest a significant amount of time, so caregiving demands can manifest as a barrier for participants. More future research on evidence-based interventions could create robust benefits for older workers with caregiving demands. Intervention studies of flexible work arrangements as well as state and federal policies and practices are warranted.

We know too little about the health and well-being of older workers, as well as of those who are not working but may wish to do so under certain conditions and those who are working unwillingly. For example, too little is known about the micro-level impacts on older adults’ work and retirement of large-scale social changes—in technology, the economy, the labor market, and society writ large. Much of what we know about the later work course in relation to health and well-being comes from studies of earlier cohorts confronting very different demographic, technological, social, and economic forces, as well as different private sector and public policy regimes. Beneficial future research could explore the contemporary—and changing—experiences of work and retirement and the conditions shaping health and well-being. Long-term trends may affect work at older ages: rising inequality in income, wealth, and health; changes in the nature of work and employer-provided benefits; changes in family structure, caregiving needs, and women’s roles; and the rising share of Hispanics in the U.S. population. Across socioeconomic status and demographic groups, the heterogeneous effects of these trends, as well as of the current COVID-19 pandemic and economic recession, demand investigation.

This page intentionally left blank.