Labor Force Exits and COVID-19: Who Left, and Are They Coming Back?

October 04, 2021

In March and April 2020, as the United States began grappling with the COVID-19 pandemic, employment fell by roughly 25 million people. Of those 25 million, some quickly returned to work, while others became unemployed or left the labor force altogether. In this blog post, we investigate the characteristics of those individuals who left their job at the beginning of the pandemic, follow up on their status a year later, and predict how many will be employed a year from now.

Who Initially Left the Labor Force?

We used the Current Population Survey, which interviews selected individuals for four consecutive months, pauses for eight months and then resumes interviewing those individuals for another four months. These CPS data allow us to follow those who had a job separation in either March or April 2020. Of those with a job separation at the beginning of the pandemic, we classified workers into three groups based on their employment status at the end of the initial four months of observation:The observation period began in February 2020 for those who separated from their jobs in March, and it started in March 2020 for those who separated in April.

  • “exiters,” those who left the labor force
  • “nonexiters,” those who remained unemployed
  • “reemployed,” those who quickly returned to work

The table below shows select demographic breakdowns for these three groups.

Characteristics of Workers with Job Separations in March or April 2020
Exiters Nonexiters Reemployed
Male 40% 48% 49%
Female 60% 52% 51%
Older than 55 38% 20% 26%
Had Been on Temporary Layoff 23% 85% 64%
Has Children Age 12 or Younger 16% 18% 23%
White 76% 74% 78%
Black 12% 10% 12%
Asian/Pacific Islander 8% 10% 5%
Other Race or Two or More Races 4% 6% 4%
Less than High School Diploma 16% 13% 12%
High School Diploma 28% 27% 27%
Some College 29% 33% 34%
Bachelor's Degree or Higher 27% 27% 27%
Composition of Job Separations 29% 25% 46%
NOTES: Characteristics are for surveyed workers who were separated from their jobs in either March or April 2020 and who could be tracked for two months after the separation. These workers were divided into: “exiters,” those who left the labor force by their final month in the CPS panel; “nonexiters,” those who remained unemployed in their final month in the panel; and “reemployed,” those who had returned to work by the end of the panel.
SOURCES: Current Population Survey and authors’ calculations.

While a majority of all three groups were female, women made up a larger fraction of exiters (60%) than of nonexiters (52%) and reemployed (51%). In addition, nonexiters and reemployed were more likely to indicate that they had been on temporary layoff following their job separation (85% and 64%, respectively, versus 23% for exiters), which means they were given a return-to-work date or expected back at the same job within the next three months. Finally, exiters had a larger population over 55 (38%) than nonexiters (20%) and reemployed (26%). This suggests that some exiters may be early retirees. Indeed, that seems to be the case: 36% of exiters between ages 55 and 64 reported being retired, and 88% of exiters age 65 and older reported being retired. Across other groupings, the composition of exiters, non-exiters and reemployed appears to be similar.

One Year Follow-up

The structure of the CPS allows us to see where the exiters and nonexiters are a year after their initial displacement. We located exiters and nonexiters who returned to the survey in either March or April of 2021 and looked at their labor force status in their final month of observation (which falls in either May or June 2021). The next table shows a breakdown of the final labor force statuses for exiters and nonexiters.

Labor Force Status of Exiters and Nonexiters, May/June 2021
Employed Unemployed Not in Labor Force
Exiters 34% 8% 58%
Nonexiters 60% 14% 27%
NOTES: The employment status of surveyed workers who were separated from their jobs in either March or April 2020 and either left the labor force (exiters) or remained unemployed (nonexiters) within two months of their original separation. Status was determined in May or June 2021.
SOURCES: Current Population Survey and authors’ calculations.

This analysis shows that a year later, about one-third of exiters had returned to employment, while nearly 60% remained out of the labor force. Given the presence of so many older workers and retirees in this group, it is not so surprising that many have not gone back to work.

Nonexiters were much more likely to resume working: 60% had returned to work, but 27% later left the labor force. Of the nonexiters who found work again, 39% of them found jobs in different industries, compared with 57% of exiters. Coupled with the high incidence of temporary unemployment among nonexiters, the fact that the majority of the nonexiters who returned to work had not switched industries suggests that many in this group returned to their old jobs.

How Many Workers Will Go Back to Work by June 2022?

Using the transition patterns of the workers we can track between 2020 and 2021, we infer that out of the roughly 25 million people who lost their jobs at the start of the pandemic, 15.6 million were back in employment, 1.7 million were unemployed, and 7.7 million were out of the labor force as of June 2021. Given that a substantial number of individuals remain out of the labor force, the next question is whether they will return any time soon.

To estimate this likelihood of returning to work, we constructed a model that takes in demographic characteristics and reasons for being nonemployed from our CPS sample of exiters and nonexiters, and calculates the probabilities that these individuals were employed in their last observed month in 2021. The next step was to extrapolate these probabilities to the rest of the nonemployment pool, which includes people beyond those who lost their jobs in March or April of 2020 and who appeared in the survey a year ago (the sample used to define exiters, nonexiters, and reemployed groups above). For this nonemployment pool as a whole, we know their demographics and reasons for being nonemployed.

We applied our model to the sample of individuals from the CPS who were observed as out of the labor force or unemployed in February 2021 but employed in February 2020 in order to get estimated probabilities of these workers being employed again within a year. This strategy takes into account the recent composition of the nonemployment pool for individuals who, on the basis of being previously employed recently, could come back to employment. The distributions of the predicted probabilities for this sample are shown in the histograms below.

NOTE: The estimated probabilities of returning to work by February 2022 are for those who lost their jobs after February 2020 and remained unemployed as of February 2021.

SOURCES: Current Population Survey and authors’ calculations.

NOTE: The estimated probabilities of returning to work by February 2022 are for those who lost their jobs after February 2020 and left the labor force as of February 2021.

SOURCES: Current Population Survey and authors’ calculations.

Based on our model, the average unemployed worker has a 54% chance of being employed again within a year, and the average out-of-labor force worker has a 35% chance of being employed again within a year. Applying these probabilities to the estimated number of exiters and nonexiters who were still out of employment in June 2021, (1.7 million unemployed and 7.7 million out of  the labor force), we expect 3.6 million more individuals to return to employment by June 2022. For perspective, in June, we were short by about 7.1 million jobs compared with the pre-pandemic level. This analysis suggests that this gap will only close about half by June 2022.

Our findings come with some caveats—it is difficult to extrapolate using 2020-21 outcomes because this was an unusual time period. There are other factors that were not present over this time period that may impact the recovery, such as the expiration of enhanced unemployment benefits, risk related to the delta variant or other new variants, and additional jobs added from new entrants into the labor force. In any case, the characteristics of the individuals who remain out of work and, in particular, their labor force attachment will play key roles in the recovery going forward.

Notes and References

1 The observation period began in February 2020 for those who separated from their jobs in March, and it started in March 2020 for those who separated in April.

About the Authors
Victoria Gregory
Victoria Gregory

Victoria Gregory is an economist with the Federal Reserve Bank of St. Louis Research division. Learn more about the author

Victoria Gregory
Victoria Gregory

Victoria Gregory is an economist with the Federal Reserve Bank of St. Louis Research division. Learn more about the author

Joel Steinberg

Joel Steinberg is a research associate at the Federal Reserve Bank of St. Louis. 

Joel Steinberg

Joel Steinberg is a research associate at the Federal Reserve Bank of St. Louis. 

This blog offers relevant commentary, analysis, research and data from our economists and other St. Louis Fed experts. Views expressed are not necessarily those of the St. Louis Fed or Federal Reserve System.


Email Us

Media questions

All other blog-related questions