Derailed by the COVID-19 Economy? An Intersectional and Life Course Analysis of Older Adults’ Shifting Work Attachments

Abstract

This paper addresses the uneven employment effects on older Americans (aged 50–75) of the COVID-19 pandemic. Drawing on monthly Current Population Survey data from January through December 2020, we take an intersectional and life course approach to study the labor market effects of COVID-19 on older Americans. First, we chart monthly labor force states throughout 2020 for older adult subgroups defined by age, gender, and race/ethnicity. We then examine transitions out of and into work from one month to the next. We find gendered age-graded declines in employment, increases in unemployment, and increases in the proportions of people in their 50s reporting they are not in the labor force for other reasons (NILF-other), most dramatically for Asian and Hispanic women. There is little change in age-graded retirement from before to during the pandemic, regardless of gender or race/ethnicity, though there are education-level effects, with those without a college degree more likely to retire in the face of COVID-19. White men with a college degree are the most apt to retain their work engagement.

Research Questions and Background

The COVID-19 pandemic and subsequent public health policies led to significant reductions in economic activity and employment opportunities, resulting in widespread unemployment (U.S. Bureau of Labor Statistics, 2020a, 2020b). In addition, many Americans moved out of the workforce entirely, often for “other” reasons than retirement, disability, or illness (Moen et al., 2020). A life course approach (Elder, 1985) emphasizes the importance of the timing of the COVID-19 disruption in people’s lives. For example, Americans in their 20s were particularly at risk of both unemployment and exiting for other reasons (Moen et al., 2020). We can also theorize distinctive and varied impacts on the work attachments of older Americans as they move to and through the conventional retirement years. How did workers in the large cohort of Boomers (born 1946–1964, aged 56–74) and the leading edge of Gen Xers (born 1965–1970, aged 50–55) fare in the face of COVID-19? The dramatic COVID-related drop in employment and the slow pace of economic recovery may have uniquely affected the later work course of older Americans, especially those historically disadvantaged in the labor market.

Both economic and health effects of the pandemic could have had labor market repercussions for older adults. Older workers may have suddenly found themselves without a job, but the type of exit they made is unclear. As Brand (2015) points out, job loss is not the same thing as unemployment. Older adults losing their jobs during the 2020 pandemic year could have defined themselves as unemployed (actively looking for work), as retired, as leaving the workforce to care for themselves or other people in their lives, or not even looking for new jobs in recognition of ageism and the low likelihood of being hired at their life stage in the face of a labor market upended by COVID-19. Older adults’ heightened health vulnerability to the coronavirus may have uniquely influenced decisions to look for new jobs given constrained labor market opportunities.

We draw on monthly nationally representative population data from the 12 months of 2020, just prior to and during the COVID-19 pandemic, to investigate labor force trends for older Americans in the midst of this health and economic crisis. We focus on four questions: What proportions of older Americans continued to work, despite the pandemic? What proportions found themselves suddenly unemployed? Did they retire in larger numbers during the pandemic compared to pre-COVID? Or did some simply drop out of the workforce for “other” reasons?

A Life Course Approach to Social Change

We draw on a gendered life course approach (Elder, 1985; Moen, 2013, 2016), modified to incorporate a feminist intersectional lens (Brewer et al., 2002; Collins, 2015; Collins & Bilge, 2020; Crenshaw, 1991; Romero, 2018) to assess COVID-19 work-related inequalities for older Americans. Our analysis focuses on gendered later life course stages as they unfold within interlocking social identities, including race and class in addition to gender and age. This framing theorizes disparate effects of social change depending on socio-historical contexts, as signaled by time period (Elder, 1985) and on-going systems of inequality (Collins, 2015; Collins & Bilge, 2020). It also underscores the importance of biographical timing (individuals’ ages when COVID-19 occurred) and the cumulative advantage and disadvantage of subgroups of the population (Dannefer, 2020; Elder, 1985).

Socio-Historical Contexts

Both life course and intersectionality approaches theorize differential and dynamic effects on later-adult work engagement depending on socio-historical contexts (Collins, 2015; Collins & Bilge, 2020; Elder, 1985), in terms of both the shifting economy across 2020 and entrenched macro-level institutionalized structures shaping employment options. COVID-19 constitutes a remarkable societal shift, an abrupt social change in labor market conditions unfolding first rapidly then more slowly each month from March through December 2020.

But there is also the context of pre-pandemic “baseline” intersectional inequalities, the socio-historical location of individuals in jointly configured categories of disparate privilege and power (Collins & Bilge, 2020). An intersectional lens suggests differential experiences by social groups based on their social-locational identities. These social inequalities can be seen in pre-pandemic January, in the uneven work states of older women and men depending on their mutually constructed combinations of age, gender, race/ethnicity, and social class. Were these existing inequities exacerbated in April, when COVID-19 had its greatest impact on the economy? What happened as the United States moved through the ensuing months of 2020?

Timing

Life course scholars key in on the timing of historical disruptions in people’s lives (Elder, 1985). Age is a marker of social as well as biographical time, with institutionalized age- and gender-graded patterns of work and retirement in later adulthood. This is evident in gendered age norms and expectations about how much and how long to work, but also in terms of ages of eligibility for programs such as Social Security and Medicare.

Because older adults in their pre-retirement 50s are more apt than those in their 60s to be in the labor force (Moen et al., 2021), we hypothesize greater employment changes due to COVID-19 for those in their 50s. This is also the group of older adults most likely to still have children at home or to be caring for older relatives (Stoiko & Strough, 2019; van der Horst et al., 2017), possibly pushing them to remain out of the workforce in the face of heightened pandemic-related care demands. By contrast, many of those in their early and later 60s are on various exit or nonwork paths (Cahill et al., 2015; Calvo et al., 2018; Moen et al., 2021) and may well have taken COVID-19 job loss as an occasion to retire or leave the workforce in some other fashion.

Cumulative Dis/Advantage

A third life course theme concerns resources, opportunities, and constraints that cumulate across the life course, perpetuating or even heightening labor market advantage or disadvantage (Dannefer, 2020). Earlier (pre-pandemic) research documents race and gender disparities in later adult employment, with older White men, especially those with a college education, working longer than other subgroups of the population. Older White women are more apt than women from other races and ethnicities to continue working into their 60s, though at lower rates than White men (Brown & Warner, 2008; Cahill et al., 2015; Calvo et al., 2018; Moen et al., 2021; Warner & Brown, 2011; Warner et al., 2010). Gendered later life courses are thus interrelated and mutually constituted with other social-locational identities (Brewer et al., 2002; Collins 2015; Collins & Bilge, 2020; Moen & Spencer, 2006; Romero, 2018). Taken together, these combined locations in the social structure reflect systemic inequities in resources, opportunities, risks, and labor market outcomes.

Older adults thus come to the COVID-19 crisis with different work attachments, different jobs, different life histories of gendered and racialized discrimination, and at different stages of their work course. They also face possibly exacerbated age discrimination, given widespread public discussions about older people’s COVID-19 health risks.

Whether the labor market effects of the pandemic are similar or different across women and men in their 50s and 60s of different races/ethnicities is an open question. Of particular interest is the dynamics of social inequality in the face of job displacement. Did previously advantaged subgroups with high work participation (Asian and White men) fare best, or did they too suddenly experience major disruptions in their employment attachments? Given job loss, did previously advantaged subgroups recover more quickly than other categories of older Americans?

Data and Methods

Data

To address these questions, we analyze monthly labor force data (January through December 2020) from the Current Population Survey (CPS) prepared by IPUMS (Flood et al., 2020). The CPS data contain large, nationally representative samples of the civilian, non-institutionalized population and detailed information about labor force participation and attachment. We limit our sample to individuals aged 50–75. CPS interviews occur each month and questions are asked about the week that contains the 12^th. As such, despite skyrocketing unemployment claims in the last 2 weeks of March 2020, large increases in unemployment are not observed until the April 2020 CPS. Our repeated cross-sectional analyses across this 12 month period illustrate heterogeneity in pandemic effects on the employment status for older adults.

We also use the panel feature of the CPS to capture micro-level month-to-month changes, as older women and men who are working one month either remain or stop working, and those not working either remain in that state or else (re)enter paid work. This analysis is possible given the participation of many CPS respondents in the survey up to eight times (in 4 months, out 8 months, and back in 4 months) over a 16-month period (Drew et al., 2014). Up to 75% of the CPS sample appear in two adjacent months.

Key Outcomes

We classify older Americans by their labor market state each month, as employed full time or part time, unemployed (actively looking for work), retired, disabled, or not in the labor for other reasons (NILF-other). Note that the proportion of people unemployed represents the fraction of the population that is unemployed. This contrasts with the official unemployment rate, which indicates the proportion of those in the labor force who are unemployed. As such, the proportion of older adults who are unemployed will be lower than the official unemployment rate for older adults. “Retired” captures the proportions of people who are not in the labor force who report the reason as being “retired,” as opposed to those who report being disabled or NILF for other reasons.

Our work transition outcomes encapsulate short-term continuity and change across each two consecutive months (for instance, March to April, or May to June) in two broad states: working and not working. “Working” combines both full-time and part-time employed individuals who are actually working that month. “Not working” consists of those who are not on the job—furloughed, temporarily laid off, unemployed, retired, disabled, and NILF-other. Using these two broad states, we examine movements of older men and women between work and nonwork from one month to the next.

Key Intersectional Social Locations

Gender is a vital social-locational marker given our gendered life course and intersectional framing, leading to our analysis of the different experiences of older women and men. Age captures stage in the life course, the timing in people’s later life course when the COVID-19 disruption occurs, as well as institutionalized age-graded norms and policies (eligibility for Social Security and Medicare). We use age as a categorical variable representing 5-year age groups between 50 and 75 (those older than 75 experienced stability in that most were retired pre-COVID). Race is coded as Black, Hispanic, White, and Asian. We omit individuals identifying as Native American or multi-racial from our intersectional analyses given small cell sizes. Socioeconomic status (SES) is based on whether individuals have a college degree. Table 1 provides a description of the nationally representative sample of older American women and men in the CPS from January through December 2020. Work status in Table 1 is for pre-pandemic January 2020 for women (N = 23,623) and men (N = 20,692).

Table 1.

Description of Sample.

	Women		Men
	Observations	Weighted percent	Observations	Weighted percent
White	2,05,622	73.2	1,84,440	74.8
Black	26,526	11.8	19,703	10.3
Hispanic	17,362	9.2	15,516	9.6
Asian	12,973	5.8	10,910	5.3
50–54	39,183	16.4	36,999	17.8
55–59	43,797	17.5	40,801	18.8
60–64	45,730	17.5	41,473	18.2
65–69	41,524	15.3	36,695	15.4
70–74	35,755	12.7	30,981	12.5
75+	56,494	20.7	43,620	17.4
No college degree	1,77,291	67.1	1,50,206	64.5
College or above	85,192	32.9	80,363	35.5
Retired*	10,731	44.3	7,963	36.7
Disabled*	1,750	7.5	1,466	7.3
NILF-other*	1,411	6.5	498	2.5
Unemployed*	217	1.0	316	1.5
Self-employed*	168	0.6	360	1.6
Part-time, economic/unknown*	244	1.1	186	0.9
Part-time, non-economic*	1,672	6.9	1,025	4.8
Full-time*	6,493	28.1	6,829	34.7
Full-time, long hours*	937	4.0	2,049	10.1
Working* (employed and at work)	9,169	39.3	10,038	50.2
Not working*	14,454	60.7	10,654	49.8

Note: * Indicates observations and weighted percent in January 2020.

Methods

Our goal is to understand the disparate effects on older Americans’ work attachments of COVID-19. To do so, we first chart proportions of older American women and men in different labor force states at different ages, both prior to and throughout the months of the pandemic in 2020. Second, we conduct an intersectional analysis by age group, gender, and race/ethnicity, identifying continuity and change in the proportions of those who are working, unemployed, retired, disabled, or not in the labor for other reasons (NILF-other). Third, we draw on panel data on older women and men across two time points to examine intra-individual transitions—from working to not working, and from not working to working across the months of 2020, considering as well the effects of educational attainment on these transitions in combination with gender, race/ethnicity, and age.

Note that this is a person-centered approach, examining subsets of people in terms of interlocking configurations of their social-locational identities, rather than looking at relative effects on employment of variables, such as the effects of age “net” of gender, or race/ethnicity “net” of social class as gauged by educational level (see also Ragin & Fiss, 2017). In this way, we can capture later life course social inequalities in the labor market that would be obscured by conventional multivariate analysis.

Repeated Cross-Sectional Results

Age/Gender Disparities in Later Adult Work

Table 2 captures the shifting pandemic-driven historical context—proportions of older American women and men in various employment states across age groups in January 2020 (pre-COVID), April 2020 (height of the pandemic downturn), and December 2020 (initial economic recovery). The numbers in January demonstrate “baseline” gender disparities in older adult work attachments, as well as the gendered, age-graded nature of work and retirement in the later life course. These inequalities are firmly entrenched before the pandemic. For instance, in January, smaller proportions of older women than older men worked full time and larger proportions retired earlier than men. COVID-19 further accentuated gender disparities; Table 2 reveals marked pandemic-driven declines in full-time employment in April, especially for women in their 50s. It also shows striking continuity in age-graded retirement trends from January through December. In other words, despite massive displacement, there was no rush by older women or men into retirement.

Table 2.

Employment States by Age Group in January, April, and December 2020, Women and Men.

Women
		50–54	55–59	60–64	65–69	70–74	75+
Retired	January	4.50	10.00	29.56	60.22	76.75	86.84
	April	4.03	9.91	27.63	61.69	77.94	87.71
	December	4.27	10.72	28.39	60.41	78.80	88.87
Disabled	January	8.32	11.74	11.06	6.57	3.71	3.61
	April	7.51	10.24	10.49	6.02	3.24	3.76
	December	7.56	10.92	10.91	6.19	3.84	2.72
NILF-other	January	12.27	10.57	8.17	4.26	2.23	2.04
	April	17.24	14.37	10.15	4.65	3.13	1.97
	December	14.46	12.25	9.02	4.77	3.26	1.80
Unemployed	January	1.55	1.72	1.40	0.70	0.50	0.21
	April	9.36	8.89	6.95	4.30	2.58	1.30
	December	4.03	3.73	2.25	1.62	0.83	0.42
Part-time	January	8.87	9.60	9.97	10.03	7.96	3.17
	April	6.56	6.77	6.78	7.20	5.47	2.33
	December	9.43	7.95	9.10	8.56	5.65	2.56
Full-time	January	64.49	56.37	39.83	18.23	8.83	4.13
	April	55.30	49.83	37.98	16.15	7.63	2.92
	December	60.25	54.43	40.34	18.45	7.63	3.64
Men
		50–54	55–59	60–64	65–69	70–74	75+
Retired	January	2.90	7.66	22.15	52.13	70.48	83.39
	April	2.60	6.55	22.08	53.40	73.83	85.96
	December	3.79	7.27	22.69	53.07	72.03	84.42
Disabled	January	6.97	10.09	12.51	7.03	3.43	2.89
	April	7.13	9.85	11.28	5.03	2.93	2.31
	December	7.40	10.49	10.83	5.92	3.40	2.40
NILF-other	January	3.30	3.80	2.54	2.16	1.96	1.37
	April	6.70	6.30	4.20	3.27	1.71	1.14
	December	4.24	5.10	4.27	2.70	2.16	1.35
Unemployed	January	2.32	2.50	1.81	1.20	0.63	0.41
	April	9.21	8.37	6.36	4.83	3.47	1.39
	December	4.16	4.11	3.58	1.88	1.01	0.53
Part-time	January	3.41	3.91	5.85	8.65	8.60	4.85
	April	3.31	3.96	5.40	6.17	5.53	3.06
	December	3.59	4.49	5.52	7.51	6.76	4.61
Full-time	January	81.11	72.04	55.15	28.83	14.89	7.09
	April	71.04	64.97	50.68	27.30	12.53	6.14
	December	76.83	68.54	53.10	28.93	14.65	6.68

Timing matters as well. Recall that because so many women and men in their 50s are working for pay, we posited that they were the group of older adults most likely to be affected by the COVID-19 economic downturn. Indeed, we find that the biggest COVID-related shifts in April occur among adults in their 50s. There were massive exits from full-time work, with sharp increases in unemployment and in being out of the labor force for reasons other than retirement or disability (NILF-other).

Women aged 50–54 experience approximately a ten-percentage point drop in in full-time work at the onset of the pandemic downturn, from 64.5% in January to 55% in April. Men in their early 50s see a similar ten-percentage point decline in working full time, from 81.1% in January to 71.07% in April. This contrasts with women and men in their early 60s; women’s already low full-time employment rate in their early 60s declines only two percentage points from January to April (39.83%–37.98%), even as full-time employment drops five percentage points (from 55.25% to 50.68) for men in this age group.

There are some reductions as well in working part time and in being out of the workforce because of a disability, especially for those in their 50s. These changes attenuate somewhat by December, but do not return to the pre-pandemic levels observed in January.

Cumulative Intersectional Dis/Advantage

Employed

Figure 1 shows monthly trends in employment (full- or part-time) for women Figure 1(a) and men Figure 1(b) in their 50s by race and ethnicity. January through March represent pre-COVID disparities in employment. Focusing first on pre-pandemic January, women in their 50s were disadvantaged in that they were less likely to be employed than men, regardless of race/ethnicity. However, there are within-gender disparities, with White women most likely, and Hispanic women the least likely, to be employed before COVID-19 disrupted the economy. White, Asian, and Hispanic men in their 50s are also the most advantaged pre-COVID in employment status, with Black men the least likely to be working Figure 1(b).

Figure 1.

Percent employed.

Women in Their 50s

The proportion employed declines by April (compared to January) for every subgroup, with one historically disadvantaged subgroup—Hispanic women in their early 50s—experiencing the greatest drop (18 percentage points), from 65 percent employed in January down to only 47 percent in April. By contrast, the smallest COVID-related employment downturn occurs for another disadvantaged subgroup, Black women in their early 50s, who experience a six-percentage point decline. White and Asian women aged 50–54 have declines of 12-percentage points. Asian women in the second half of their 50s also undergo a 12-percentage point decline in the proportion employed, while Black women aged 55–59 have only a 5-percentage point decline. There is a slight upward trend toward the end of the year for women in their early and later 50s, though 50-something women remain employed at lower rates in December than January, regardless of race and ethnicity.

Men in Their 50s

In contrast to Black women, Black men in their early 50s experience the largest percentage point decline for men in this age group, going from 76% employed in January to 64% in April, dropping further to 58% in June, a total drop of 18 percentage points Figure 1(b). Hispanic men are also heavily impacted, dropping 16 percentage points from 87% in January to 71% in April. Thus, two historically disadvantaged subgroups of men in their 50s experience cumulative disadvantage with COVID-19.

There is also suggestive evidence of disadvantage among previously advantaged Asians, though small numbers suggest cautious interpretation. We find Asian men aged 55–59 experience a 25-percentage point drop (from 87% in January to 62% by June). The proportion of Black men working in the latter half of their 50s declines by eight percentage points during this period. Note the age-graded nature of employment; for example, 76% of Black men aged 50–54 are employed in pre-COVID January, compared to only 69% of Black men aged 55–59.

Women in Their 60s

A different picture emerges for women in their 60s. Figure 1(c) reveals less pre-pandemic social inequality by race/ethnicity in proportions of women in their early 60s employed. Around half of women in this age group are employed in January, varying only from 49% of Hispanic women to 52% of Asian women. But subsequent rebound is not equally distributed. There is a suggestion of some improvement in employment by December for Asian and White women aged 60 to 64, with White women’s employment slightly exceeding pre-pandemic rates (52% vs. 50%). By July, 51% and 48% of Asian and White in women in this age group are employed, respectively, compared to 41% of Black and 42% of Hispanic women in their early 60s, respectively, reflecting cumulation of advantage and disadvantage among these historically disadvantaged subgroups. From July through December, the proportion of 60–64 year old women who are employed is characterized by stability. By contrast, the recovery rates for the employment of Black and Hispanic women in their early 60s by December are slower than those of other women in this age group.

Looking at those 65–69 prior to the pandemic, Hispanic women were the least likely to be employed (21% compared to 28% of Asian and Black women and 30% of White women in this age group). Small race/ethnicity disparities in the employment rates of women in their late 60s remain by December (ranging from 24% for Black and Hispanic women to 27% for White women and 30% for Asian women).

Men in Their 60s

Black men in the first half of their 60s are the least likely of race/ethnic subgroups to be employed both before (49% in January) and during (41% in May) the pandemic, a decline of 8 percentage points. This does not improve as the economy started recovering. By December, only 43% of Black men in their early 60s are working.

In contrast, pre-COVID-19, over six in ten Asian, Hispanic, and White men aged 60–64 are employed. At the onset of the pandemic downturn, White men in their early 60s experience little change in employment, compared to Hispanic and Asian men in their early 60s. In fact, employment rates of White men remain roughly flat; 63% employed in January, 59% in April, and 61% employed in December. Asian men in their early 60s return to employment quickly following large pandemic-driven declines, with Hispanic men in their early 60s not rebounding as well.

Turning to men in the second half of their 60s we find lower pre-pandemic employment rates overall (ranging from 30% for Black men to 45% for Asian men). Black men aged 65–69 have the lowest rates of employment for this age group before and during COVID-19: 30% in January dropping to 24% by April and remaining low (23%) in December. Employment for Hispanic men aged 65–69 drops as well, and also remains low, at 34% in December. White men in their later 60s experience more stability; 38% employed in both January and December, with only a 5% decline (to 33%) in April.

Unemployed

Women in Their 50s

Figure 2(a) shows the sharp increase from low unemployment pre-COVID to high unemployment for women in their 50s due to the pandemic downturn. The proportions of women unemployed January through March are small, especially for non-White women, so results should be interpreted cautiously. For women in their early 50s, unemployment is highest in April among women of color. Black women’s unemployment remains highest (at 6% in December), whereas White women’s unemployment is only three percent by the end of the year.

Figure 2.

Percent unemployed.

For women in their late 50s, Asian and Hispanic women experience the highest unemployment with peaks in April and recovery by December. Black and White women in their late 50s have largely the same changes in the proportion unemployed for most of 2020, going from 3% and 1% in January to 9% and 8% in April, respectively. The proportion unemployed for Black women in their late 50s starts trending up at the end of 2020 while for White women, it drops to 3%.

Men in Their 50s

As was the case for women, unemployment results should be interpreted cautiously given small numbers of men of color in the data who are unemployed each month when the data are stratified by age. White men aged 50–54 experience peak unemployment at 8% in April (from 2% in January), dropping back near pre-pandemic rate of 3% in December. Higher proportions of men of color tend to be unemployed during each month of the pandemic and continue to be more likely to experience unemployment in December. White men aged 55–59 have a steady downward trend in their unemployment rates after April. By contrast, Hispanic, Black, and Asian men in this age group have upward trends in unemployment to finish 2020.

Women and Men in Their 60s

Pre-pandemic unemployment is low for women and men in their 60s regardless of race or ethnicity. Unemployment increases for women in all subgroups in their early 60s by April. It then drops for Hispanic, Black, and White women, but remains high for Asian women (at 5%) in December. Black and White men in the second half of their 60s report the lowest proportions unemployed in the middle of COVID (8% for White men in April, 9% for Black men in May). Racial gaps in unemployment among men in their later 60s attenuate by the end of 2020, closing the gaps between Asian and Hispanic men and their Black and White counterparts.

Not in the Workforce because of Retirement, Disability, or “Other” Reasons

Despite evidence of alternative pathways and widening disparities around retirement timing (Brown & Warner, 2008; Cahill et al., 2015; Calvo et al., 2018; Moen et al., 2021; Warner & Brown, 2011; Warner et al., 2010), we find clear age- and gender-graded pre-pandemic patterns but few racial/ethnic divergences in the percentages of women and men in their 50s and 60s describing themselves as retired (Figure 3). Neither do we find patterned increases in retirement after the onset of the pandemic. Instead, we see some declines in the proportion retired in the second half of 2020 among Asian women aged 65–69 and Asian men aged 60–64. Moreover, there are few pandemic-related increases in reports of not working because of a disability, with Black women and men most likely to have a disability-related exit both before and during the pandemic, regardless of the month considered (data available but not shown).

Figure 3.

Percent retired.

There are COVID-related age, gender, and race/ethnic intersectional effects in the proportions not in the labor force for “other” reasons, with one in five Hispanic and Asian women in their early 50s reporting they are NILF-other in May/June Figure 4(a). These two subgroups retain high levels of NILF-other across 2020, with nearly one in four (24%) Hispanic women and one in five (19%) Asian women aged 50–54 and one in five Hispanic women (19%) and one in four Asian women in the second half of their 50s doing so in December. Similar but lower rates for these subgroups occur for those in their early 60s. Previous studies have shown that women in their 50s and 60s are likely to be in this category because of family caregiving responsibilities (Moen et al., 2020; Moen et al., 2021). Black, Asian, and Hispanic men in their early and late 50s also have higher COVID-related NILF-other rates but this may reflect low pre-COVID-19 levels.

Figure 4.

Percent NILF for other reasons.

Social Class Limiting COVID-19 Effects

Figure 5 shows the intersectional effects of age, gender, race/ethnicity, and education (no college degree/college degree) on proportions working (employed and at work) from January through December. Here, we are examining individuals employed and at work, given that many employed during 2020 were furloughed or otherwise not actually in a job due to the pandemic. These data demonstrate the value of a college education, both before and during the pandemic, for women and men across race and ethnicity categories. There are greater racial/ethnic disparities in proportions working among women and men in their 50s without, compared to those with, a college degree. Asian women without a college degree experience the greatest COVID-19-related displacement, but also the greatest rebound by December.

Figure 5.

Percent working by educational attainment.

Transitions In and Out of Work

The repeated cross-sectional trends reported thus far are enormously informative, but do not capture the dynamics of work transitions that older Americans may endure—or choose—during COVID-19. Accordingly, we examine month-to-month shifts in whether different categories of workers exit their jobs, as well as whether those not working take on paid work the following month (Figures 6 and 7). We consider the lived intersectional experiences of women and men in their 50s and 60s, first by race/ethnicity and then by race/ethnicity in combination with educational attainment. We focus here on the rate that individuals transition to and from working (employed and at work) to highlight the dynamic disruptions caused by the COVID-19 pandemic.

Figure 6.

Transition rates between working (W) and not working (N) by edu. attainment, aged 50–59.

Figure 7.

Transition rates between working (W) and not working (N) by edu. attainment, aged 60–69.

Exiting Work: The Advantage of a College Degree

Having a college degree lowers the risk of leaving work for both women and men in their 50s and 60s of all races and ethnicities. But there are within-education disparities. The transition rates from being on the job to no longer working from March to April for those with no college degree are high for all subgroups of 50-something women and men, but highest for non-college-educated Asian women and men. Nearly half of Asian women in their 50s with no college degree are at work in March but no longer working by April. This is also the case for about one in three less-educated Asian men in their 50s.

By contrast, college-educated women and men in their 50s experience fewer transitions out of work. Least apt to transition into no longer working are educated White and Asian men in their 50s, those most attached to the workforce pre-pandemic.

Moving Back on the Job

Figures 6(c) and (d) and 7(c) and (d) show the complex dynamics of returning to work across 2020. The proportions of women in their 50s moving (back) into working are highest in June, July, and August, with no clear race/ethnic disparities in moving back into work. Few women in their 60s who are not at work the preceding month transition into work. Men in their 50s begin moving back into work in greater proportions from May onward, with Black men in their 50s and 60s the least apt to do so.

Opportunities for returning to work are stratified by education. Transitions into paid work are highest in May, June, and July for college-educated Hispanic women in their 60s, and for college-educated Hispanic men in this age group by July. Still, those without a college degree also transition (back) into working, but at lower rates. Some upticks in returning are already apparent by May for Asian, Hispanic, and White men in their 60s without a college degree.

Discussion

We draw on nationally representative population data and three key life course themes—socio-historical context, timing, and cumulative dis/advantage—overlaid by a feminist intersectional lens—to capture disparities in the later work courses of older Americans given the pandemic derailment in paid work. We tease out patterned COVID-19 impacts on social inequalities in older Americans’ employment states from pre-pandemic January 2020 through its height in April and on through December 2020.

Socio-Historical Contexts

COVID-19 itself constitutes a massive socio-historical disruption that played out over 2020, with differential employment effects from April through December. Our contribution is that we chart monthly shifts as they unfold across the 2020 pandemic year, not only in employment and unemployment, but in other work states (such as retirement and being out of the workforce for other reasons) as they evolve in different ways for different categories of individuals.

Our evidence shows that April was indeed the cruelest month in terms of heightened unemployment and reduced employment. But its impacts continue over the year for some, with remarkable rebound for others. For example, many older women remain out of the workforce for “other” reasons throughout 2020, most typically to provide care for family members. But significant portions of women and men in their 50s and 60s are not working and not retired—simply not in the workforce. Future research might fruitfully consider how people having different exits fare as the economy recovers more fully.

Timing Matters

Older adults confront the COVID-19 pandemic and its labor market fallout in the context of work attachments that are already age-as well as gender-graded, with 50-something men the most apt to be employed, women in their 50s and 60s the most apt to exit work at earlier ages, and women and men in their 70s almost all retired. These age and gender trends have timing implications for the labor market impacts of COVID-19, touching the work lives of older Americans in their 50s far more than those in their 60s (or 70s). There are clearly selection effects, in that many in their 60s and 70s have already exited the workforce through retirement prior to COVID-19.

We show that the COVID pandemic did not precipitate large-scale moves into retirement. Although much scholarship on this age group equates not working with being retired, older Americans offer different reasons for being out of the workforce beyond retirement, including because of a disability or for “other” reasons (Moen et al., 2021; see also Warner et al., 2010). Surprising to us is the absence of much pandemic-driven change—and few race/ethnicity effects—in the proportions reporting being retired or disabled among women and men in their 50s and 60s.

Our analysis demonstrates the importance of looking at age categories, not simply “controlling” for age. The pandemic alters the employment attachments of 50- and early 60-something women and men in a variety of ways—unemployment and non-employment, including exiting because of family care. Research is needed on who left the workforce in the face of COVID-19, whether they were forced out or chose to leave, and their subsequent experiences.

Cumulative Dis/Advantage

For both women and men, COVID-19 constitutes an unprecedented derailment in their later work courses after March 2020. But these downturns are unevenly distributed, reflecting both cumulative dis/advantage and emergent inequalities. This is no simple story of White advantage and non-White disadvantage. To be sure, White men experience less displacement and more stability than change in their employment across 2020.

But our evidence is more complicated. Consider, for example, that with the onset of COVID-19, Black men in their early 50s, already with lower rates of employment, typically have lower rates of pandemic-driven employment compared to both Hispanic and Asian men. And Hispanic women, already less likely to be working than White and Asian women prior to the pandemic, have this disparity further accentuated by COVID-related downturns, suggesting processes of cumulative disadvantage.

Black women and men in their 50s have different experiences, with Black women having less of a decline in employment, even as Black and Hispanic men have large declines. The similar rates of displacement of Black and Hispanic men are noteworthy, in that Hispanic men are far more likely to be employed prior to the pandemic.

The effects on Asian women and men have to be treated with caution, given small sample sizes when divided by age. Asian women and men in their 50s could be considered “advantaged” in that they have high rates of employment before the pandemic. But this declines precipitously by June 2020. Asian and Hispanic men in their 60s are also advantaged in terms of high pre-COVID-19 employment participation, and yet our evidence is suggestive that Hispanic and Asian men in their early 60s are especially hard hit. Still, Asian as well as White men experience high rates of rebound by December, possibly reflecting their advantaged position in the labor market.

In the months prior to the pandemic, Asian and Hispanic women were more likely than White or Black women to be out of the workforce for “other” (not disability, not retirement) reasons, and this accelerates in tandem with the pandemic, most probably to care for children, grandchildren, or older relatives. Asian and Black men in their late 50s and early 60s also report somewhat of a COVID-driven increase in being NILF-other, but not to the extent of women in this age group, and most likely not for care-providing reasons (Moen et al., 2020, 2021). Clearly, simply focusing on unemployment fails to capture the labor market costs of the pandemic disturbance. More in-depth analysis of specific subgroups of the older population could illuminate their distinctive experiences.

The Value of Education

We document higher education as a key resource in mitigating job displacement effects of the COVID-19 downturn and in hastening subsequent rebound by December. In fact, having a college degree suppresses many of the racial/ethnic disparities in proportions employed for women and men in their 50s. At the other end of the spectrum is the persistent disadvantage of those without higher education. Black and Hispanic women and men in their 50s with no college degree experience the highest rates of work displacement.

We also assess the value of a college degree in shaping the pandemic’s effect on short-term transitions. Looking at the group of older Americans working in March who were no longer working by April reveals important class-driven impacts of COVID-19, with less-educated men and women much more likely to stop working. Similar class (education) effects are seen in those who subsequently move from not working to working during the beginning the recovery period in June and July; those with a college degree move back into employment at higher rates than those without a college degree. These protective effects of higher education operate within as well as across race/ethnic disparities.

There are also combined class and racial/ethnic differences in the effects of COVID-19 for people who were working in March but displaced by April. Asian and Hispanic women without a college degree and similarly less-educated Asian and Black men in their 50s are the most likely to experience pandemic-driven transitions out of work. Scholarship is needed on the interplay between educational level and the types of occupations and industries most apt to weather the pandemic.

Conclusions

Older Americans came to this pandemic year in the historical context of an already uneven playing field producing different life histories in terms of their previous work courses, resources, and constraints. Even before the pandemic, women and men in the vast Boomer cohort and the edge of the GenX cohort in their 50s and 60s were embedded within intersectional axes of disadvantage. In looking at the effects of COVID-19 on economic and social outcomes, it is important to consider this existing “baseline” context of disadvantage. Our goal of capturing the complexity of COVID-19 effects across mutually shaping social-locational identities precipitated our use of a gendered life course approach in combination with an intersectional lens as analytic tools (Collins & Bilge, 2020).

The fact is, older adults enter 2020 with existing disparities in risks and opportunity structures, including systemic age-, race- and gender-discrimination and class-based employment options. For example, in pre-pandemic January, Hispanic women and Black men in their 50s and 60s are on average less likely to be employed compared to other subgroups of the same age. These contexts and conditions are interactive, not additive, pointing to the need to assess patterned descriptions of mutually constituted categories of people rather than using multiple regression analyses to examine variables, not people, such of the effects of race “net” of gender or age (see also Ragin & Fiss, 2017).

The complexity of what we report is the hallmark of an intersectional approach (Brewer et al., 2002; Collins & Bilge, 2020; Romero, 2018), and, in turn, points to the importance of considering layers of systemic advantage and disadvantage (Dannefer, 2020) as they converge in the lives of individuals of different ages, genders, races/ethnicities, and social classes. What we show is not merely cumulative advantage or disadvantage in labor force derailments intensified by COVID-19, but, rather, nuanced intricacy. For instance, there is suggestive evidence that previously “advantaged” older Asian women and men are especially at risk during the pandemic, along with their less-advantaged Black counterparts. And having a college degree meant greater employment protection for those in their 50s and 60s of all races and ethnicities, likely indicating jobs with more flexibility (enabling the move to working from home accompanying shutdowns) and lower likelihood of working in service occupations associated with greater risk of job loss.

Disparities in the labor market among older Americans also point to the importance of future research on the stratification of historically advantaged and disadvantaged subgroups in particular occupations and industries. A key mechanism producing differential labor market effects is the fact that the COVID-19-induced contraction affected some sectors of the labor market more than others, producing unequal risks of layoffs. For example, the pandemic dealt a particularly hard blow to the retail, hospitality, and leisure sectors, but strengthened the technology sector. Unequal distribution of older workers across different occupations and industries may well account for much of the disparate COVID-19 effects we document. Also needed is deeper analysis of differential health-precipitated exits, along with the effects of organizational and public policy environments shaping later adult employment possibilities in the time of COVID-19 (Kelly & Moen, 2021).

What is also required in future scholarship is what CPS data cannot provide, further understanding of the meanings of COVID-19 job dislocations in the lives of individuals. This is a real limitation of our investigation. Future analysis using different sources of data is required to capture the psychosocial impacts of pandemic-driven exits. For example, do conventional findings on the emotional costs of unemployment (Brand, 2015) hold in times of massive social disruption? This is key to understanding the human meanings of COVID-19 derailments in paid work as well as of other disruptions. Clearly, more research is required on the disparities and dynamics of Americans of all ages working—or not—in the contexts of the volatile social, economic, and technological—as well as public health—changes characterizing our times.

Author Biographies

Phyllis Moen directs the Life Course Center on the Demography and Economics of Aging, holds a McKnight Presidential Endowed Chair, and is professor of Sociology at the University of Minnesota. Her latest (award-winning) books are Encore Adulthood: Boomers on the Edge of Risk, Renewal, and Purpose (Moen, 2016) and Overload: How Good Jobs Went Bad and What We Can Do About It (Kelly & Moen, 2021).

Joseph H. Pedtke is an Economics PhD student at the University of Minnesota. His research focuses on topics such as inequality, structural change, education and child investment, and macroeconomics. He has published work on labor market dynamics across demographic groups, racial disparities in economic outcomes, and historical patterns around financial crises.

Sarah Flood is Director of U.S. Survey Projects at IPUMS at the University of Minnesota, affiliate of the Minnesota Population Center, and Associate Director of the Life Course Center. She oversees data infrastructure projects funded by the National Institutes of Health to develop, support, and improve access to U.S. population data. She has published work on the time use and well-being of parents, couples, and older adults.