3

Work and Retirement Pathways

In this chapter, we offer a conceptual framework to depict and understand the patterned work and retirement pathways of older adults. A key feature of this conceptual framework is its emphasis on the existence of multiple pathways for older adults to continue working, as well as to transition between working and nonwork states. This feature is aligned with the continuous trend for workers to move away from the traditional linear career progression over the past 30 years (e.g., see Figure 2-7 in this report for the increasing trend of self-employment from 2004 to 2019; also see Calvo et al., 2018; McDonough et al., 2017; Wang et al., 2013; Johnson et al., 2010). As such, instead of viewing full-time retirement as a one-time event and an inevitable end of one’s workforce participation, we acknowledge that at least some individuals can remain active in various forms of workforce participation in later adult life and that such workforce participation can be flexible in nature until the very end of one’s life (e.g., Moen, 2016a; Zhan et al., 2015; Wang et al., 2008).

On the basis of this conceptual framework, we then review the main theories that have been used to explain work and retirement pathways for older adults. Further, we consider the proximal forces that shape older adults’ work and retirement pathways, which include the preferences, expectations, and experienced constraints regarding their workforce participation. We review research findings regarding these forces, especially on the effects of their various empirical referents and variables. Toward the end of this chapter, we review findings regarding disparities and heterogeneity in work and retirement pathways. Finally, we discuss challenges faced by this research area and make recommendations for future research.

A CONCEPTUAL FRAMEWORK OF WORK AND RETIREMENT PATHWAYS

In retirement research, retirement pathways typically refers to courses of action that people follow to exit the workforce (Szinovacz, 2003; Flippen and Tienda, 2000). Traditionally, and especially before mandatory retirement was abolished in the United States, the course of action was viewed as quite simple, containing a single transition from full-time employment to full-time retirement (Beehr, 1986). However, with changes in policies, work, and workforce, the courses of action for exiting the workforce have evolved to contain multiple possibilities.

Accordingly, retirement has been defined in various ways in the literature (Denton and Spencer, 2009), including:

• as nonparticipation in the labor force;
• as reduction in hours worked and/or earnings;

Following our conceptual framework, an older adult’s work and retirement pathway can be measured as their work/nonwork states sequenced over time. As such, this conceptual framework captures two important realities. First, in terms of tracing a person’s workforce participation, it captures that an older adult can transition in and out of work over time. This matches the empirical findings from the first five waves of the HRS, which showed three workforce participation patterns after retirement from career jobs (Wang and Chan, 2011). The three patterns include two “stayer” classes, one containing retirees who were employed at each point in time after retirement and one containing retirees who were never employed at any point after retirement; and one “mover” class, which includes employees who transition between employment and retirement, back and forth. Other studies also found wide variability in the timing of “unretirement,” that is, moving back into the workforce following an exit from one’s career job (Moen et al., 2021; Calvo et al., 2018; McDonough et al., 2017; Cahill et al., 2016, 2015).

The second reality captured by the framework concerns individual differences in workforce participation over time: work and retirement pathways differ among older adults. Thus, this framework offers a way to conceptualize the heterogeneity in work and retirement pathways among older adults. This again matches the findings from Wang and Chan (2011) and others (Moen et al., 2021; Calvo et al., 2018; McDonough et al., 2017), which show that the years of education were a key predictor of retirees’ workforce participation patterns. Specifically, retirees who had more years of education are less likely to be either “stayers who were never employed” or “stayers who were always employed” than to be “movers.”

THEORETICAL MECHANISMS

In this section, we review the main theoretical approaches that have been used to explain work and retirement pathways for older adults. They are listed in Table 3-1.

Rational Choice Theory

Rational choice theory explains how older workers’ financial resources and the external economic environment are related to their retirement decisions (Laitner and Sonnega, 2013). It considers the retirement decision to be the result of utility maximization. In particular, financial resources (both current and expected future) affect retirement decisions through their relationship to (current and expected future) consumption, and this relationship is affected by work decisions and saving and consumption decisions. Workers will retire when the utility they obtain from not working is higher than from continuing to work, which is more likely to occur when they place greater value on leisure time (nonwork) and when they have accumulated sufficient financial resources to finance their preferred consumption level in retirement. This depends on many things, including their actual and predicted health, expected returns on their assets, and more. They are more likely to continue to work when the utility of leisure time is lower, and when working longer will enable them to generate sufficiently higher utility from consumption after retirement to compensate for foregoing leisure (retirement) earlier.

TABLE 3-1 Theoretical Approaches

Thus, when applying rational choice theory to consider retirement and workforce participation options, the concept of subjective expected utility is very useful. It guides researchers to consider older workers’ subjective preferences and expectations regarding their workforce participation options. Accordingly, utility-maximizing models of labor supply, human capital investment, consumption, and retirement are useful in predicting and explaining people’s workforce exit pathways (Lazear, 1986; Hanoch and Honig, 1983; Blinder and Weiss, 1976).

Theory of Planned Behavior

The theory of planned behavior (Ajzen, 1991; depicted in Figure 3-4) explains workers’ workforce exit decisions by considering their attitudes toward their jobs, employers, careers, and retirement, as well as the norms in their workplace (e.g., Zhan et al., 2013; Shultz, 2003; Adams and Beehr, 1998) and perceived ability to control the behavior (that is, to make the choice consequential). This expectancy-value theory highlights the importance of workers’ attitudes toward retirement, that is, whether they evaluate it favorably or unfavorably, and toward its alternative (continuing to work) in influencing their workforce exit decisions. It also emphasizes the role of perceived social influence, that is, the subjective norm—the belief about whether most people approve or disapprove of the behavior—regarding retiring in affecting an individual’s workforce exit decision. Finally, perceived behavior control, that is, a person’s perception of the ease or difficulty of performing the behavior of interest, is closely linked to constraints experienced by older workers when making workforce exit decisions. Together, these three dimensions—expected value, subjective norms, and perceived control about work and retirement—powerfully predict an individual’s intentions, which in turn reliably predicts their planned behavior.

Role Theory

Role theory (Ashforth, 2001) focuses on how global identity derived from a social position with an expected behavioral repertoire shapes behavior. Once individuals adopt a specific role, such as a work role, their behaviors and decisions are influenced by the activated role identity. Role theory conceptualizes nonwork states as being free of work-role identity and considers the process of moving from employment to nonemployment as a role-based

Preferred Workforce Exit Pathways

In light of these factors, preferences as to when to exit the workforce differ across subgroups of the population, with women, those with less education, and those in physically demanding and low-status occupations more likely to prefer and take early exits, even as other psychosocial factors matter as well (Calvo et al., 2013; Zhan et al., 2013; Davies and Cartwright, 2011; Wang and Shultz, 2010; Wang et al., 2008; Barnes-Farrell, 2003). Preferences as to retirement timing do, in fact, predict actual timing of retirement and actual exits, although many find themselves “retired” before they expected as a result of poor health, caregiving obligations, buyouts, or layoffs (Moen, 2016a). One Swedish study finds that personal preferences for retirement timing have little effect on actual retirement timing net of other factors (e.g., health and other circumstances described in Chapter 4), suggesting that opportunities and constraints may outweigh subjective preferences in shaping pathways to retirement (Örestig et al., 2013). Further, preferences for working or retiring can be clouded by a sense of ambivalence for those who are not sure what retirement will bring, as well as for those who would like to work “some,” but in less demanding, more meaningful, and more flexible ways but see no such options (Moen, 2007).

As women’s labor market attachments have increased, decision-making and experiences around retirement timing have become more likely to be couple-based, not individual-based processes (Henkens and van Solinge, 2002). Older couples can confront disjunctures in both partners’ work/nonwork preferences and behavior. Moreover, spouses may have preferences for their husbands’ and wives’ exits as well as their own. For example, a person’s leisure may be more valuable if the spouse also retires, because the spouse’s leisure time can augment the subjective utility of the person’s leisure.

In the past, given their more variable attachments to the labor force, women have tended to tailor their work exits around those of their husbands, but this is changing as women begin to remain in the workforce throughout the life course, even when their children are young (Moen et al., 2006). A study using Swedish register data finds that for about one in four couples, both individuals retire within a year, and that women who synchronize retirement timing with their husbands tend to retire at earlier ages, even as men who synchronize with their wives are more likely to delay, thus retiring at later ages (Gustafson, 2017). A recent study of the oldest members of the Baby Boom cohort (using the HRS) shows that men influence their wives’ retirement timing, even as women appear to have little influence on their husband’s retirement timing (Stancanelli, 2017; Jackson, 2016), although these findings are not consistent with those of another study (also using the HRS) examining an earlier cohort (Coile, 2004), which found the opposite: men are responsive to their wife’s retirement incentives, but not vice versa. Other factors such as maintaining access to employer health insurance may also affect spouses’ retirement timing (Boyle and Lahey, 2016; Witman, 2015; see also Chapter 4).

Preferred Job and Workplace Characteristics

For some older adults, continued work in later life contributes to the fulfillment of psychological needs such as “the need to belong, to be a contributing member of society,” and to stay engaged (James et al., 2016, p. 334; see James et al., 2020). In addition to greater financial security (and perhaps employee benefits), continued work in later life can confer a sense of belonging, self-esteem, identity continuity, and a way to structure time (Zhan et al., 2019; 2015; Smyer and Pitt-Catsouphes, 2007). The Special Committee on Aging report from the U.S. Senate (2017) concluded that work is also linked to improved mental and physical health and overall quality of life, although empirically this can be hard to distinguish from causality running in the other direction—with healthy and well people working longer. Chapter 5 includes a detailed discussion regarding elements of job and workplace characteristics that older workers prefer.

Expectations

People’s expectations often determine their choices, but these processes are again a mix of thoughtful deliberation and less systematic beliefs. People’s ability to forecast their future happiness is flawed; in a variety of life decisions, comparing expectations versus reality, affective forecasting is poor (Gilbert and Wilson, 2007). People’s expectations about their work-exit pathway also include finances and work context.

Financial Considerations

One important expectation that accompanies remaining in the workforce is that work will provide better financial income. A person may expect to receive more income by working longer than the typical retirement age (Cahill et al., 2013), but this may not be the sole consideration. A job that meets such financial expectations may breach other expectations about engaging in meaningful work if there is a lack of person-job fit and the job is less intrinsically rewarding. Consistent with the distinction between push and pull factors of retirement (Wang et al., 2009; Shultz et al., 1998), retirees who are “pulled” into bridge employment due to positive expectations about work activities may have higher levels of retirement satisfaction than retirees who are “pushed” into bridge employment due to financial expectations.

Work Context Shaping Expectations

When making workforce exit decisions, older workers’ expectations are likely shaped by factors in their work context. These include not only retirement-related organizational-level policies and norms, but also perceived biases and negative stereotypes that older workers encounter in their workplace or in society more broadly (Wang and Shultz, 2010). In particular, retirement age norms in the workplace and in society play important roles in shaping older workers’ retirement expectations and plans (Settersten and Hagestad, 1996). Individuals who are behind normative career advancement schedules or have plateaued in their careers are less likely to expect investment and fair treatment from their organizations. Accordingly, they are more likely to feel pressure from the organization and society to retire.

Workplace policies and practices related to the treatment of older workers may also be critical in forming age-discrimination-related expectations that may influence older workers’ attachment to organization and retirement intention (Boehm et al., 2014; Goldberg et al., 2013). For example, data from a sample of German companies showed that age-inclusive human resource practices (e.g., equal opportunities to training and promotion for all age groups) promoted the age diversity climate, leading older workers to expect fair treatment and positive social exchange with their supervisors, which enhanced their intentions to remain and work for the same organization (Boehm et al., 2014).

Finally, the workplace as a social community also provides important interpersonal context to expectations about the benefits of working longer. Research on employee retention has emphasized the role of interpersonal links in reducing employees’ workforce turnover behaviors (Lee et al., 2004). The impact of interpersonal context might be stronger for older employees, as older adults value social goals such as maintaining and building satisfying social relationships to a greater extent than younger ones (Carstensen, 1995). For example, having a dense and strong friendship network in the workplace may enhance older workers’ sense of social belonging and acceptance, likely to fulfill expectations of positive social experiences for continuing to work and thus serving as a key contextual factor in retaining older workers.

Constraints

Constraints for Work and Retirement Pathways

In addition to preferences and expectations, constraints also shape workforce exit pathways for older workers. Economic and health resources and shocks, such as changes in Social Security or Medicare policy, chronic strains, such as burnout and precarity, and family caregiving responsibilities and the circumstances of spousal work and income all influence both preferences for and the actual timing of exits (Fast et al., 2020; Keating et al., 2019; Stoilko and Strough, 2019; Kalleberg, 2018; Gustafson, 2017; Beehr and Feldman 2011; Raymo and Sweeney, 2006; Moen et al., 2006; Dentinger and Clarkberg, 2002). So too does a sense of perceived “workability,” the subjective sense that one is able to do the job, which is related to health, a sense of control, and job demands (McGonagle et al., 2015). Related to the last, disability trends from 1997–2010 show increased disabilities for those ages 40–64 (Martin, Schoeni et al., 2010), leading to increased workforce exits (Warner et al., 2010).

Early evidence from CPS data (Flood et al., 2021; Moen et al., 2020) showed that COVID-19 caused significant increases in unemployment and being out of the workforce at all ages, but especially among women and those with less education. Looking specifically at older adults in their 50s and 60s (Flood et al., 2021) shows that those in historically disadvantaged locations, in terms of race and education as well as gender, were more apt to become unemployed or leave the labor force; there was no large upturn in retirement through December 2020. In contrast to the experiences of Black and Hispanic women and men, White men’s employment was less disrupted and more apt to rebound during 2020 (see also Chapter 2). The COVID-19 pandemic significantly challenged marginalized older adults’ prospects for working longer, especially among those with lower educational attainment.

There are other forces shaping work and retirement pathways as well: technological changes coupled with corresponding skill obsolescence, encouraged (forced) early retirement packages, societal as well as organizational norms and pressures, and limited phased or alternative work options, such as possibilities for bridge jobs or reduced or flexible hours, often come into play (Hess et al., 2018; Moen, 2016; Sweet and Moen, 2012). For example, studies using HRS data of the initial wave of the Baby Boom cohort show that they are apt to take alternative work options at later ages than earlier cohorts, that poor health leads to bridge jobs (Cahill et al., 2019), and that those reaching retirement ages during the Great Recession are more apt to experience layoffs (Cahill et al., 2015). A study in the Netherlands showed that some older workers are especially vulnerable to involuntary early exits (Visser et al., 2018).

Retirement and social welfare policies shape incentives and possibilities for early or late exits, as do educational and retraining options (Foster and Walker, 2015; Zhan and Wang, 2015b; Powers and Neumark, 2003). Taking various options also shapes older workers’ relative propensity to return to work, with those taking the early Social Security option at age 62 having higher odds of subsequently doing part-time work (Kail and Warner, 2013). As discussed in Chapter 5, working conditions, such as greater control over the time and place of work, can also color expectations of older workers for the future, such as their expectations about working past certain ages (Moen et al., 2016; Pienta and Hayward, 2002).

As described in socioemotional selectivity theory, older adults selectively invest their time and energy resources as they recognize their time horizons are constrained (Carstensen, 2011). They also seek to exercise control over life transitions (Staudinger et al., 2016; Heckhausen and Schulz, 1995). Voluntary work and voluntary exits promote well-being (Rhee et al., 2015; Glavin, 2013; Calvo et al., 2009; Gallo et al., 2006; Kim and Moen, 2002). But institutionalized constraints around the clockworks of work (workdays, workweeks, work years, work lives) together with macroeconomic forces limit older adults’ options for achieving their work and retirement preferences across the later life course (Moen et al., 2016; Moen, 2013).

In fact, as a life course approach underscores, paid work is organized in age-graded ways (Moen, 2016b; Settersten and Mayer, 1997). Young adults in their 20s and 30s are recruited for entry-level jobs, and occupational paths are designed for prime-age workers. Most organizational practices and pension policies are predicated on continuous full-time work throughout adulthood, followed by a one-time, one-way exit to full-time retirement, what Blau and Shvydko (2011) call labor market rigidities. This means that alternative paths—to unretire or, if working, to work less, more flexibly, and differently—are not institutionalized, rendering preferences for such options at odds with opportunities (Moen, 2016). The social organization of the work course can also foster situations like job lock, the problem that emerges when, because of Medicare health insurance eligibility, U.S. workers who would like to or need to scale back or exit the workforce are forced to continue working for the health insurance until they reach age 65 (Fisher, Ryan et al., 2016; Boyle and Lahey, 2010).

Taken together, constraints in the forms of personal, organizational, and societal circumstances and exigencies can create a mismatch between preferences and actual behavior around workforce exit-related transitions, leading to involuntary work and involuntary retirement, in contrast to similar but voluntary states (Sohier, 2019; Hershey and Henkens, 2014; Sweet and Moen, 2012; Kim and Moen, 2001). Research documents different types of mismatch, with older workers who would prefer to retire and others who would like to keep working beyond conventional retirement ages feeling unable to do so because of disability, caregiving demands, discrimination, automation, layoffs, workplace norms, or intense and rigid working conditions (Katz and Krueger, 2019; Ebbinghaus and Radl, 2015; van Solinge and Henkens, 2007).

There are those who are not working and would prefer to be working but find few if any opportunities for reemployment compatible with their skills and flexibility needs. There are also those who would like to retire but have little choice but to keep working (Sohier, 2019). These involuntary states can have harmful consequences for older workers’ well-being (Sohier, 2019; Munnell and Sass, 2008; van Solinge and Henkens, 2007; Kim and Moen, 2002).

Involuntary Dislocation from Work

The evidence from both the United States and Europe shows that many later exits from the workforce occur despite adult workers’ preference to continue working (Rhee et al., 2015; van Solinge and Henkens, 2007). These exits may take place because individuals are pushed out (Ebbinghaus and Radl, 2015; Reynolds and Wenger, 2010), experiencing health problems (Martin, Freedman et al., 2010), or serving as care providers for elderly parents, spouses, children, or grandchildren (Stoiko and Strough, 2019; van der Horst et al., 2017). Assumptions about a customary and voluntary work-to-retirement experience are clearly obsolete.

Indeed, although older workers in large numbers have reported an interest in continuing work past conventional retirement ages, there is a question about the extent to which jobs will be available to them. In the 2017 Retirement Confidence Survey¹ of American workers, only 27 percent of individuals ages 65 and older report that they have actually been able to continue working as long as planned (Greenwald et al., 2017). Chronic illnesses (self or family) often force people to work into retirement age, in order to maintain health insurance, but for this same reason, people may not be able to work as long as they need to (Homaie Rad et al., 2017).

Further, as Heidkamp and Van Horn (2020, p. 337) note,

Using data from the U.S. Bureau of Labor Statistics, for the Employment Data Digest of AARP, Schramm (2018) reports that in February 2018 the long-term unemployment rate for these older job seekers was 27.2 percent, compared to 19.0 percent for prime-age workers. Audit studies examining the effect of age on employer callbacks for unemployed workers have found employers are less likely to contact those ages 50 and over (Farber et al., 2019; Farber et al., 2016).

Thus, older job seekers must make difficult adjustments. The longer one is in a job search, the greater the stigma of unemployment, both in the United States (Liu et al., 2014; Kroft et al., 2013) and in Germany (Nüß, 2018). Older workers are typically unfamiliar with current job search techniques using social media, technology, and networking, and they often lack the training or skills for the jobs that exist (Heidkamp and Van Horn, 2020). Additionally, the growing gig economy is often less appealing to workers who are seeking stable jobs (Heidkamp and Van Horn, 2020). Together, this means that many older adults who become unemployed end up underemployed, if they find work at all, and without access to retirement benefits, paid leave, or health care (if they are under 65).

Involuntary dislocation from work is associated with an increased risk of negative health outcomes for older workers in the United States, including increased risk of cardiac disease and stroke (Gallo et al., 2006). Similarly, a study conducted in Sweden using linked employer-employees data reveals that involuntary job loss increases the risk of hospitalization, not due to strokes but to alcohol-related conditions, among both men and women, and to traffic accidents and self-harm, among men only (Eliason and Storrie, 2009). These authors report that the overall mortality risk among men increased by 44 percent during the first four years after job loss, but did not increase for women. For both sexes there was an increase in suicides and alcohol-related mortality.

It is important to note that methodologically, studying involuntary dislocations and involuntary early retirement helps researchers overcome problems of reverse causality (e.g., in measuring the influence of unemployment or

Most obvious is the cumulation of advantage with education for older White men; those who had attended college are far more likely to continue working long hours (50 or more per week) as they move to and through the conventional retirement years. Comparing women to men, women in their 50s and 60s with at least some college follow long-hour and full-time pathways at similar levels as less-educated men (those with only a high school degree or less).

Employment in the form of part-time work appears to be a pathway for well-educated women and men in their 60s (Moen et al., 2021). The probabilities of participation in this pathway are slightly higher for White women than for other women and for White men in their 60s compared to same-age men of color. There is divergence in the probability of participation in the part-time work pathway in the 60s and 70s, with the highest educated White men having higher odds of pursuing it. Older women of Hispanic or Asian/Pacific Islander background are most likely to follow the unemployed/other pathway, most typically leaving the workforce because of family-care reasons.

Gender-based Heterogeneity in Bridge Employment Decision Making

According to social gender role theory, gender roles are formed as a result of people’s social beliefs about gender differences in psychological and behavioral characteristics (Eagly et al., 2000). Gender roles in turn cultivate real gender differences in behaviors (Konrad et al., 2000). In other words, though such beliefs do not have actual roots in biological differences, they can be learned and internalized by people through socialization and impact people’s cognitions and behaviors (Eddleston et al., 2006). Given that social gender roles are associated with various work attributes, men and women may react differently to specific aspects of work (Konrad et al., 2000).

For example, people tend to expect men to pay more attention to status, prestige, and recognition that are ascribed by others. This is because striving to elicit the immediate attention of others is an important masculine role characteristic (Williams and Best, 1990). Accordingly, both meta-analytical review (Konrad et al., 2000) and empirical studies (e.g., Eddleston et al., 2006) show that men value the status aspect of a job more than women do. On the other hand, people tend to expect women to be more active in seeking positive personal relationships (i.e., “affiliation”; Williams and Best, 1990). Based on this gender characteristic, women are expected to prefer working with people rather than working alone. As such, the desire for social communion is expected to be more salient for women than men in making bridge employment decisions (Zhan et al., 2015). However, longitudinal data from a representative sample from China, controlling for age, health, education, finance, family context, preretirement job conditions, and retirement attitudes, showed that communion striving and generativity striving were positively related to bridge employment participation for both male and female retirees. However, status striving was positively related to bridge employment participation for male retirees but not for female retirees. Taken together, these findings suggest that retirees’ actual decisions to work after retirement were at least partly shaped by different needs associated with their social gender roles.

Summary

In summary, the demography of and processes leading to later adult work and retirement pathways paint a picture of enduring inequalities by historically disadvantaged subgroups who are less apt to have either resources or opportunities to work longer. The evidence to date points to the need for an intersectional approach (Lahey and Oxley 2021; Collins and Bilge 2020; Romero, 2018) to examine the effects of overlapping social statuses (such as gender, race/ethnicity, nativity as well as age), in order to promote understanding of disparities in work pathways, including the inability to exit work because of the absence of savings or a pension (see also Chapter 4) or the inability to keep working because of poor health (see also Chapter 4), working conditions (see also Chapter 5) or discrimination (see also Chapter 6). Intersectional analysis could also provide a window into the distinctive retirement preferences and experiences, as well as the gaps between them, among different subgroups of older individuals and couples. For example, this would permit a focus on the ways racism and sexism experienced early in the life course play out in the later life course, even as reports of agism increase (Gee et al., 2019).

CHALLENGES AND FUTURE RESEARCH DIRECTIONS

One challenge to understanding the dynamics and disparities in later adult work and retirement patterns is that scholars use different definitions of retirement and, accordingly, different measures (Ekerdt and DeViney, 1990). For example, “work” is often defined as full-time, long-tenure (career) employment, more typical of white-collar and unionized blue-collar White men than of those in other subgroups. However, because older adults follow multiple pathways to working in later life, it is necessary to capture specific features of work activities, such as work hours, the nature of employment, and the employment relationship.

A second challenge is that research often considers single transitions, such as from full-time work to retirement (with nonemployment equated with retirement; Radl, 2013), or from retirement to reemployment (Kail and Warner, 2013). But considering transitions this way cannot quantify the patterning or range of routes to eventual retirement, ignoring the variety of pathways (see Figure 3-3 above) followed from work to complete retirement. Recent innovations in sequence analysis, as well as panel data, are resulting in scholarship that captures the varied and unequal later life course paths older individuals experience in the United States and Europe (Moen et al., 2021; Calvo et al., 2018; McDonough et al., 2017; Fasang, 2010). Longitudinal analytic techniques can also help better depict and predict qualitative changes over time (Wang and Chan, 2011).

A third challenge involves the samples of individuals studied. Social research typically looks in the rear-view mirror, drawing on available data, often decades old, from earlier populations of older adults. This is a key issue in times of rapid social change, when the nature of both work and retirement are in flux, even as traditional safety nets are eroding (Moen, 2016). It also points to the complexity of the relationship between age, period, and cohort (Elder and George, 2016; Alwin and McCammon, 2003). Age reflects career and life course stages, but also locates people in particular socio-historical contexts. Past studies on the work and retirement pathways of earlier populations of older workers who faced very different organizational and economic environments may not apply to the experiences of contemporary older workers. Moreover, the timing of external events in the later life course matter. Consider, for example, the different experiences of Americans in the face of the COVID-19 lockdown. Older Americans in their 70s were mostly already out of the workforce, whereas those in their 50s were more vulnerable to job loss (Flood et al., 2021). This points to the need for more rapid collection and dissemination of data on contemporary older workers, as well as attention to the timing of social changes in people’s lives. It also suggests that extrapolating future trends from the experiences of recent cohorts should only be done with caution.

Relatedly, many work- and retirement-related phenomena are fundamentally intertwined with their specific socioeconomic and policy contexts (Costa, 1998). Changes have often occurred in these contexts over time, such as economic booms due to technology innovation, Social Security reform, health care reforms, and early retirement incentive practices (Wang, 2015). These changes can alter the research questions investigators ask. For example, at the end of the 1990s and in the early 2000s, much effort was invested in studying early retirement, because there was a clear trend of workers retiring earlier and earlier. However, the population aging trend, in combination with the 2008–2010 economic recession, seems to have reversed this phenomenon, and more and more older workers are choosing to postpone their retirement and stay in the workforce. As such, bridge employment becomes a focal research topic for investigation. This type of shift in research phenomena is challenging and rewarding at the same time, as new issues surface constantly.

In addition, given the context changes, the same phenomena are likely to be influenced by different factors at different points in time, which often facilitates theoretical and methodological innovations in tackling those issues. Given the disruption caused by the COVID-19 pandemic to public health, social interaction, and employment attachments, it is crucial to understand the effects of COVID-19 on subsequent work and retirement pathways.

Finally, research is still lacking regarding independent contracting and self-employment as workforce participation activities for older adults. The growth of the gig economy raises the question of the extent to which older workers are attracted to independent contractor work or are being pushed toward it. In addition, given the difficulties of measuring these populations (discussed in Chapter 2), what is the prevalence of older workers among the self-employed, and how has it changed over time? Understanding these questions could be helpful in providing a fuller picture of the employment experience of older workers. HRS data provide some insights into