Early Evidence on the Impact of Coronavirus Disease 2019 (COVID-19) and the Recession on Older Workers.

The coronavirus disease 2019 (COVID-19) pandemic and the policies to curb its spread brought parts of the U.S. economy to a virtual halt in March 2020. Payroll jobs declined by 0.7 million in March 2020 and 20.5 million in April 2020, pushing the seasonally adjusted official unemployment rate (U-3) to a peak of 14.7% in April 2020 (U.S. Bureau of Labor Statistics, 2020a, 2020b). This rate far surpasses the peak rates during the Great Recession of 2007 to 2009 (10.6%, January 2010) and the early 1980s recession (11.4%, January 1983; U.S. Bureau of Labor Statistics, 2020c), and yet the unemployment rate during COVID-19 has likely been underestimated (Montenovo et al., 2020).

Analyses of how the COVID-19 pandemic and resulting recession impact older workers has been limited. In this report, we discuss how older workers fared in prior recessions in the United States, estimate some early effects of the COVID-19 pandemic and recession on employment and unemployment rates by age group and sex, discuss how COVID-19 and this recession differ from prior recessions, and conclude with a brief discussion of important topics for future research.

How Did Older Workers Fare in Previous Recessions?

Prior research shows that older workers are differentially affected by recessions. One way they are less impacted is that in prior recessions, the probability of displacement generally declined with age, as older workers often benefitted from employment seniority (Johnson & Butrica, 2012). Older people also faced smaller increases in their unemployment rate, although that is partly because older people are more likely to leave the labor force (the unemployment rate only counts those who are actively searching for work).

In other ways, recessions hit older workers harder. Older workers took longer to find work during and after the Great Recession (Neumark & Button, 2014), and unemployed workers in their 50s also faced steeper wage losses than younger workers (Johnson & Butrica, 2012). Age discrimination in hiring also increases during recessions (Dahl & Knepper, 2020), contributing to the longer unemployment durations for older workers. Age discrimination in hiring is a significant barrier for older workers, who often rely upon temporary jobs—called bridge jobs—to delay retirement (Neumark et al., 2019).

Regardless of whether older workers face a more significant negative economic shock, any financial shock faced close to retirement could have major consequences. Coile et al. (2014) found that workers approaching retirement during a recession are disproportionately likely to suffer long-lasting negative consequences, including years of reduced replacement earnings, loss of health-care coverage, lower utilization of health care, and reduced longevity. The number of households “at risk” for being unable to maintain their preretirement standard of living in retirement also increased during the Great Recession (Munnell & Rutledge, 2013). This decrease in retirement security was driven both by increases in early retirement (Rutledge & Coe, 2012), which spreads retirement savings and Social Security over a longer time frame, and reductions in retirement assets (e.g., decreases in 401k plans).

Data and Methodology

We use data from the Current Population Survey (CPS), via IPUMS-CPS, to provide an early look (up to April 2020) at impacts on the employment outcomes of older workers. However, classifications of the employment status of individuals have been prone to errors during the COVID-19 pandemic. Therefore, we calculated two versions of the employment rate and three versions of the unemployment rate, as detailed in Bui, Button, and Picciotti (2020).

Results: Employment and Labor Force Participation Rates

Figure 1 presents employment rates by age group (ages 25–44, 55–64, and 65+ years) and sex from January 1976 to April 2020. Table 1 summarizes the changes in all the employment and unemployment rates from their average in 2019 to April 2020. Figure 1 shows a substantial decrease in employment at the early onset of the COVID-19 pandemic, a clear outlier compared to the last five recessions (shaded in gray). Figure 1 and Table 1 show that this decrease was the most substantial in percentage point terms for ages 25–44 years: a 10.92 percentage point decrease in April 2020 compared to the average in 2019. The percentage point decreases are smaller for ages 55–64 years, at 7.36 percentage points, and 65+ years, at 3.55 percentage points.

Figure 1.

Employment rates by age, January 1976 to April 2020. Authors’ calculations using data from the Current Population Survey (monthly) from 1976 to April 2020, via IPUMS-CPS. These estimates are not seasonally adjusted. The shaded areas represent official recession dates from the National Bureau of Economic Research (https://www.nber.org/cycles/cyclesmain.html). See also the notes to Table 1.

Table 1.

Changes in Employment Statistics by Age Group and Sex, April 2020

	Ages 25–44			Ages 55–64			Ages 65+
	Men	Women	Combined	Men	Women	Combined	Men	Women	Combined
Employment rate, %
Average rate in 2019	87.07	73.97	80.46	70.04	58.38	64.00	24.59	16.36	20.05
Rate in April 2020	75.76	63.41	69.54	62.89	50.84	56.64	20.43	13.29	16.50
Percentage point change	–11.31	–10.56	–10.92	–7.15	–7.54	–7.36	–4.16	–3.07	–3.55
Percent change, %	–12.99	–14.28	–13.58	–10.21	–12.92	–11.49	–16.92	–18.77	–17.71
Employment rate, excluding “not at work last week,” %
Average rate in 2019	85.09	70.99	77.98	67.60	55.85	61.51	23.36	15.44	18.99
Rate in April 2020	70.60	57.67	64.08	57.90	45.71	51.58	17.64	11.33	14.17
Percentage point change	–14.49	–13.32	–13.90	–9.70	–10.14	–9.93	–5.72	–4.11	–4.82
Percent change, %	–17.03	–18.76	–17.83	–14.35	–18.16	–16.14	–24.49	–26.62	–25.38
Labor force participation rate, %
Average rate in 2019	89.96	76.56	83.20	71.84	59.97	65.68	25.32	16.87	20.66
Rate in April 2020	86.41	73.54	79.92	70.60	59.04	64.61	23.82	15.99	19.51
Percentage point change	–3.55	–3.02	–3.28	–1.24	–0.93	–1.07	–1.50	–0.88	–1.15
Percent change	–3.95	–3.94	–3.94	–1.73	–1.55	–1.63	–5.92	–5.22	–5.57
Official, U-3 unemployment rate, %
Average rate in 2019	3.22	3.38	3.29	2.50	2.65	2.57	2.90	3.04	2.96
Rate in April 2020	12.32	13.77	12.99	10.92	13.89	12.33	14.26	16.85	15.43
Percentage point change	9.10	10.39	9.70	8.42	11.24	9.76	11.36	13.81	12.47
Percent change	282.61	307.40	294.83	336.80	424.15	379.77	391.72	453.28	421.28
U-6 unemployment rate, %
Average rate in 2019	7.26	8.40	7.79	6.49	7.46	6.95	10.60	12.50	11.46
Rate in April 2020	22.67	24.95	23.73	20.24	23.98	22.02	26.23	29.18	27.57
Percentage point change	15.41	16.55	15.94	13.75	16.52	15.07	15.63	16.68	16.11
Percent change	212.26	197.02	204.62	211.86	221.45	216.83	147.45	133.44	140.58
“U-6+” unemployment rate, %
Average rate in 2019	8.17	9.73	8.90	7.50	8.85	8.14	11.57	13.76	12.56
Rate in April 2020	23.32	25.49	24.33	21.01	24.74	22.78	26.43	29.68	27.91
Percentage point change	15.15	15.76	15.43	13.51	15.89	14.64	14.86	15.92	15.35
Percent change	185.43	161.97	173.37	180.13	179.55	179.85	128.44	115.70	122.21

Note: Authors’ calculations using data from the Current Population Survey (monthly) from 1976 to April 2020, via IPUMS-CPS. Our “U-6+” rate is a broadened U-6 rate, as detailed in Bui, Button, & Picciotti (2020). We use the average rate in 2019 rather than the rate in April 2019 to reduce sampling variation. All estimates are not seasonally adjusted. All estimates are weighted using population weights.

While the smaller percentage point decrease in employment for older workers could be interpreted as COVID-19 having a more adverse effect on younger workers, percentage point changes are proportional to the existing employment rate. Therefore, those with higher rates would mechanically have a larger percentage point change.

Calculating percentage changes instead of percentage point changes shows that workers ages 65+ years, namely women 65+ years, had the largest percent change in employment rates. The average employment rate for women ages 65+ years was 16.36% in 2019, and this decreased by 3.07 percentage points, to 13.29%, in April 2020. This was an 18.77% decrease in the employment rate, the largest decrease for all age and sex groups (for men 65+ years it was a 16.92% decrease). One way to conceptualize this is that out of 100 women of ages 65+ years who were employed, about 19 became non-employed. The percent decrease in the employment rate was smaller for ages 25–44 years, a 13.58% decrease, and for ages 55–64 years, an 11.49% decrease. Our alternative measure of employment (Table 1) provides similar results.

There is a simple interpretation of this discrepancy between larger percentage point decreases in the employment rates of younger workers and men, but larger percent decreases in the employment rates of older workers and women. The economic shock of COVID-19 hit employed individuals of ages 65+ years and women more harshly, but while a smaller economic shock hit younger workers and men, there were more of them that were impacted due to higher existing employment rates, hence the higher percentage point decreases.

Figure 2 and Table 1 present labor force participation rates. There were no significant decreases in the participation rates in previous recessions for any group, but there has been a clear decrease since 2019 for all age groups. Like with the employment rate, those ages 65+ years experienced the largest percent decrease (5.57%). Unfortunately, the data do not allow us to track retirement; however, Coibion, Gorodnichenko, and Weber’s (2020) analysis of the Nielsen Homescan panel data suggests that retirement increased significantly. Before COVID-19, 52.7% of those not working or looking for work stated that they were retired, and this increased to 59.5% in early April 2020.

Figure 2.

Labor force participation rates by age, January 1976 to April 2020. See the notes to Table 1 and Figure 1. These estimates are not seasonally adjusted.

Results: Unemployment Rates

Figure 3 and Table 1 present the official (U-3) unemployment rates. The increase in the unemployment rate in April 2020 was dramatic and rapid, with unemployment rates increasing by a staggering 13.81 percentage points for women ages 65+ years, compared to the average rates in 2019, resulting in an unemployment rate of 16.85%. For men ages 65+ years this was an 11.36 percentage point increase, resulting in an unemployment rate of 14.26%. Unemployment rates also increased significantly, but by much less, for the younger age groups: 9.70 percentage point increase for ages 25–44 years versus 9.76 for ages 55–64 years. These were massive percent changes in unemployment rates, between 282% and 454%. Workers ages 65+ years and women faced both higher percentage point and percent increases in unemployment rates.

Figure 3.

Official (U-3) unemployment rates by age, January 1976 to April 2020. See the notes to Table 1 and Figure 1. These estimates are not seasonally adjusted.

The increase in the unemployment rate in April 2020 was dramatic and rapid, with unemployment rates increasing by a staggering 13.81 percentage points for women ages 65+ years, compared to the average rates in 2019, resulting in an unemployment rate of 16.85%. For men ages 65+ years this was an 11.36 percentage point increase, resulting in an unemployment rate of 14.26%.

This rapid and severe increase in the unemployment rate has been much more dramatic than the increases in previous recessions. In Figure 3, the increase in the unemployment rate in the early onset of COVID-19 was two to three times larger than that from the peak of the Great Recession of 2007–2009. We also see that the COVID-19 recession impacted workers ages 65+ years more than younger workers, while this was the opposite in previous recessions.

Figure 4 presents the broader U-6 measure of unemployment rates by age group and sex, and Table 1 summarizes the percentage point and percent decreases in the U-6 rate. The gap between the U-6 and U-3 rates was the largest for workers ages 65+ years, especially for women, meaning that they were more likely to be marginally attached to the labor force or underemployed (these groups are counted in the U-6 rate). Using the U-6 rate shows an even higher unemployment rate in April 2020, of between 20.24 and 29.18%. The percentage point increase in the U-6 rate was again the largest for ages 65+ years, especially for women, but the difference in the percentage point increase by age group and sex was less pronounced using the U-6 rate. Results using our further broadened “U-6+” rate are similar (Table 1).

Figure 4.

U-6 Unemployment rates by age, January 1994 to April 2020. See the notes to Table 1 and Figure 1. These estimates are not seasonally adjusted.

For the U-3 rate, we find that workers ages 65+ years faced both significantly larger percentage point and percent increases in unemployment rates, but with the U-6 rate we find that workers ages 65+ years faced only slightly higher percentage point increases, but lower percent increases. This discrepancy suggests that the impacts on younger workers were more muted, as these workers were more likely, compared to workers ages 65+ years, to move from employed to underemployed instead of from employed to non-employed.

Why Might the Effects of the COVID-19 Pandemic and Recession be Different?

Our early data suggest that COVID-19 hit the economy hard, increasing unemployment rates to levels not seen since the Great Depression. We find that older workers, especially women, were more impacted by COVID-19 and this recession than by previous recessions, where older workers were generally less affected. The fact that older workers were hit harder this time is likely driven by the higher mortality rates for older people. For those that got COVID-19, mortality rates were 8.6% for those in their 70s, 4% for those in their 60s, 1.25% for those in their 50s, and 0.3% for those in their 40s (Begley, 2020). As of June 17, 2020, 92.6% of COVID-19 deaths in the United States were in those ages 55 years and older (National Center for Health Statistics, 2020).

Another unique feature about COVID-19 and this recession is that it is now more difficult for older workers to work longer to compensate for the large decrease in their retirement savings (e.g., 401k plan balances fell about 19% in the first quarter of 2020; Brandus, 2020). First, working longer is less safe. This is especially the case for many bridge jobs that older workers take to delay retirement. Examples of common bridge jobs include managerial positions, transportation drivers, sales, and construction roles for men, and managerial, administrative/clerical, sales, and personal service roles for women (Cahill et al., 2011; Neumark et al., 2019). However, bridge jobs appear to often have higher rates of face-to-face contact and are less likely to allow for remote work, especially for women (Montenovo et al., 2020). Early evidence suggests that retirement increased significantly in the first month of the pandemic (Coibion et al., 2020), suggesting that older workers are retiring early instead of taking the risk.

Even if older people brave the elevated risks of applying for new jobs, they may face even more age discrimination in addition to the high levels already occurring, especially during recessions (Dahl & Knepper, 2020; Neumark & Button, 2014). Employers may, for example, assume that older people are riskier to employ because their age group is statistically more susceptible to COVID-19. These barriers to working longer will increase early retirement rates, Social Security claims, and poverty among retirees.

Conclusion

Older workers are often hit hard by recessions, and often harder than younger workers. Recessions cause reductions in employment and earnings, increased early Social Security claiming, reduced retirement savings, and increased poverty at old ages. This is due both to the COVID-19 recession having a larger magnitude, but also because the pandemic creates additional risks for older workers that cut their (work) lives short.

The COVID-19 pandemic and resulting recession hit those near retirement ages, especially women, much harder than past recessions, and much harder than younger groups.

There are several policy implications of our findings. First, older people will face even worse retirement security, given the decreases in retirement savings accounts and other assets and the reduced ability to work longer. Many older people will also retire earlier than they would like, which essentially reduces their monthly income by spreading Social Security and other savings over a longer time frame. Older women are disproportionately harmed because they have lower retirement security and benefit much more from working longer (Maestas, 2018), which is even less possible during the pandemic.

Governments should consider programs to provide more income support to older people, such as increasing Social Security benefits or expanding job training or employment programs, such as the Senior Community Service Employment Program. This support is likely to be especially important for older women, given that they already face worse retirement security, have a larger benefit to working longer, and face more age discrimination (Neumark et al., 2019). Increased mortality from COVID-19 also makes older women more likely to become widowed. This increases poverty among older women, because widowed women face high poverty rates (Burn et al., 2020).

Second, the large increase in early retirement (as suggested by Coibion et al., 2020) will further strain the Social Security Trust Funds through early Social Security benefits claiming. The Social Security Trust Funds were already in a difficult position before COVID-19 as they were forecasted to be exhausted by 2035 (Brandus, 2020). The federal government may need to enact reforms to improve the solvency of the system earlier than anticipated. These reforms could include cutting benefits, such as by further increasing the full benefits retirement age beyond 67 years. However, this would further decrease retirement security. A better option may be to increase revenue for the system by increasing Social Security taxes, borrowing more money, or cutting expenditures elsewhere in the federal budget to maintain or increase Social Security benefits.

There is significant room for future research. We plan to update our analysis as new data become available, including by exploring more intersectional impacts (e.g., race and age). Future work can also explore data that were not available to us. Data from the Health and Retirement Study are the most useful for studying older workers, although they are released with a long lag. For example, future work could use these data to analyze bridge jobs (e.g., Cahill et al., 2011) or effects on retirement and labor supply (e.g., Rutledge & Coe, 2012).