View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

Home / Publications / Research / Economic Brief / 2021

Economic Brief
February 2021, No. 21-06

High Labor Market Churn During the 2020 Recession
Article by: Claudia Macaluso

Download Data Appendix

Richmond Fed research has found that job losses during the COVID-19 recession
have been concentrated in high-turnover sectors, with turnover rates in those
occupations even higher than they were during the Great Recession. Workers
displaced from high-turnover occupations often avoid long periods of
unemployment, but they are historically less likely to develop long-term
employment relationships, which limits their potential for sustained wage
One way the current pandemic-induced recession stands apart from previous recessions is
the high level of labor market churn. Loosely speaking, labor market churn refers to the
pace of reallocation of workers and jobs, that is, the magnitude of job creation and job
destruction ows alongside the size of hiring and separation ows. A high-churn labor
market is characterized by workers cycling through di erent jobs and between employment
and nonemployment at a high pace.1 This Economic Brief shows that the labor market in
the COVID-19 recession displayed a higher level of churn than in the Great Recession of
2007–09. It also discusses the interpretation of these data and the implications for policy.
Data on labor market ows clearly show that job losses during the COVID-19 recession have
been concentrated in high-turnover sectors and that turnover rates in these occupations
have been even higher than during the Great Recession. The faster pace of worker
reallocation across occupations is both a blessing and a curse: On one hand, workers
displaced from high-turnover occupations are less likely to experience persistent
nonemployment spells because they often become reemployed in sectors that use similar
skills. For example, workers previously employed in the personal care services sector often
transition into sales occupations and vice versa. On the other hand, workers employed in
these occupations are historically less likely to develop long-term employment

relationships, which limits their potential for sustained wage growth in the medium run and
long run. The increased pace of separations and turnover during the COVID-19 recession
may further exacerbate this limitation.

The Flow Approach to Labor Markets
Some labor market indicators are merely snapshots of what the labor market looks like at
any point in time — for example, the employment-to-population ratio, the labor force
participation rate or the unemployment rate. Other approaches to labor market analysis
focus on changes, rather than levels, in the employment and unemployment rates with the
goal of depicting labor market dynamics. This type of analysis is commonly referred to as
the ow approach to labor markets. This approach captures the frequency at which U.S.
workers cycle across jobs and between employment and nonemployment, providing a
more nuanced perspective on employment in ows and out ows.
Current Population Survey (CPS) data allow researchers to estimate average monthly ows
of individuals among employment, unemployment and the labor force, as well as between
jobs. A seminal 2006 article by Steven J. Davis of the University of Chicago, R. Jason
Faberman of the Chicago Fed and John C. Haltiwanger of the University of Maryland
pioneered the interpretation of these ows.2 Job creation and job destruction are job ows,
while hires and separations are worker ows. The sum of worker ows and job ows is
called "total labor market churn." These ows are quite large: In an average month, we see
122 million employed people, 2.8 million people who change jobs, 1.4 million who become
unemployed and 3 million who exit the labor force from employment. Similarly, we have 6.2
million unemployed people, 1.8 million who become employed and 1.4 million who exit the
labor force from unemployment. Finally, in an average month, we have 59.3 million people
who are not in the labor force, 1.4 million who join the labor force as unemployed (that is,
they begin actively seeking work) and 2.8 million who join the labor force by accepting a job.
These numbers indicate the high uidity of the U.S. labor market, which has been linked to
the exceptional resilience of the American economy during previous downturns.3
The scale of job ows and worker ows also varies widely by industry. For example, job- ow
rates are approximately twice as large in leisure and hospitality — the industry in which
most food and maintenance workers are employed — than in manufacturing. Worker ow
rates are approximately three times larger for leisure and hospitality than for

Adding Occupational Flows to the Picture
In addition to ows among states of the labor market (employment, unemployment and
out of the labor force), ows between occupations are also important. Consider, for
instance, a worker who loses her job as an administrative assistant and takes a job — out of


necessity — as a restaurant server. This transition to such a highly di erent occupation
entails "starting from scratch" in developing occupation-speci c human capital. Intuitively,
we would expect greater earnings losses from a transition to a dissimilar occupation. This
intuition is borne out by the data.4
In fact, research suggests that occupational ows can be exceptionally important in
determining future earnings prospects and human capital accumulation. And during
recessions, occupational transitions can account for the majority of negative displacement
e ects. Lisa B. Kahn of the University of Rochester examined the consequences of
graduating from college in a "bad" economy in a 2010 article and found a negative e ect on
occupational attainment as well as on wages.5 Even after the economy improves, these
workers are unable to fully shift into better jobs. Similarly, Davis and Till von Wachter of
UCLA concluded in a 2011 article that individuals entering the labor market during weak
economic conditions face lower earnings for many years.6 While a number of mechanisms
play a role in driving these negative e ects, the lack of opportunities for new labor market
entrants to ow into high-prestige, high-wage occupations during recessions is an
important factor.
These articles and other similar research suggest two main conclusions: 1) The American
labor market is marked by a high pace of worker ows and occupational transitions, which
results in a much lower unemployment rate than other rich countries; and 2) labor market
transitions and occupational ows are key determinants of individual outcomes, such as
employment stability, wage growth and human capital accumulation, especially during
economic downturns.

Labor Market Flows During the Pandemic
The changes in labor markets during the pandemic have presented new challenges to
researchers. Over the past year, unemployment rates have been especially turbulent,
shooting up between February and April 2020 and then declining suddenly and quickly
despite lockdowns and new waves of infections. Professional forecasters have consistently
overestimated the unemployment rate in every month since April. In recent research,
Ayşegül Şahin and Jin Yan of the University of Texas and Murat Tasci of the Cleveland Fed
have shown the value of the ow approach in such conditions, developing a ow-based
method for forecasting U.S. unemployment during the COVID-19 recession; their method
has more accurately predicted peak unemployment at the onset of the recession and its
decline throughout the year.7
Recent research at the Richmond Fed builds on the ow approach and its application to the
current recession by addressing three questions:
1. From which occupations are workers more likely to exit the labor force, and how likely
are they to reenter it?


2. How was the occupational makeup of ows into unemployment di erent in 2020 than
during the Great Recession?
3. What are the most common transitions for workers formerly employed in the hardest-hit
occupations, such as personal care, maintenance and cleaning, and food preparation
and service?
The data describing ows of workers across di erent labor market states (out of labor
force, unemployed; and employed) and across the 22 major occupational groups indicate
the following answers to these three questions. (Data on the COVID-19 recession
incorporate all of the months currently available, namely March 2020 through November
2020. For detailed results, see the data appendix .)

From which occupations are workers more likely to exit the labor force, and how likely are they to
reenter it?
The pace of transitions in and out of the labor force is somewhat unusual in the current
recession. However, this is not because individuals who are not in the labor force are less
likely to reenter it — indeed, the percentage of people who stay out of the labor force in any
two consecutive months is very similar to previous periods. In terms of ows into the labor
force, the three top occupations are o ce/administration (0.9 percent), sales (0.7 percent)
and food (0.7 percent). These numbers are very similar to the Great Recession.
One big di erence between the current downturn and the Great Recession is a much
higher ow from employment into out of the labor force. (See Figure 1.) This is especially
true for food (7.1 percent), personal care (6.1 percent) and maintenance/cleaning (5.5
percent). These job categories primarily employ women.8 In the Great Recession, these
occupations also constituted the top "donors" to the out-of-labor-force pool but with much
smaller percentages. Food was at 4.5 percent, personal care at 5.4 percent and
maintenance/cleaning at 4.2 percent.


These hard-hit occupations experience higher churn rates at all times, both in and out of
the labor force, but the pace of worker ows is even higher during the current downturn. In
the COVID-19 recession, 3.4 percent of production workers (most of them in manufacturing)
became unemployed every month, and 2.9 percent exited the labor force; in the Great
Recession, these percentages were 2.4 percent and 2.1 percent, respectively.

How was the occupational makeup of ows into unemployment di erent in 2020 than during the
Great Recession?
While persistent unemployment in the COVID-19 recession is below the level in the Great
Recession, temporary (fewer than four weeks) ows into unemployment are higher.9 This
highlights the short, but repeated, nature of job loss in the COVID-19 recession: The same
categories of workers su er high job loss month after month. For example, CPS data show
that in rst few months of the COVID-19 recession, between 3 percent and 6 percent of the
workforce experienced repeated job-loss episodes — de ned as two transitions from
employment to nonemployment within four consecutive months. (See Table 1.) For hard-hit
occupations, the share of repeatedly displaced workers is higher, with a peak of 9 percent.
In the Great Recession, this share varied between 1 percent and 3 percent for all
occupations. In the current recession, most of the higher ows into unemployment are
found in just three occupational groups: food, personal care and maintenance/cleaning (6.1


percent, 5.3 percent and 4.2 percent, respectively). (See Figure 1 above.) This contrasts with
the Great Recession both in magnitude ( ows are now about twice as high) and
occupational concentration. In the Great Recession, the only sector that stood out for its
contribution to unemployment was construction.

In any given month from February 2020 through November 2020, the fraction of
unemployed workers who remained unemployed for two consecutive months was 50.9
percent. This number is quite a bit lower than in the Great Recession (57.9 percent). Also,
ows into unemployment were lower during the Great Recession — 0.9 percent of teachers,
1.5 percent of sales workers, 1.7 percent of personal care workers and 2.3 percent of food
workers became unemployed every month. These rates have been 3.2 percent, 2.9 percent,
5.3 percent and 6.1 percent, respectively, in the COVID-19 recession.
In addition to the hardest-hit job categories, it is worth noting that the level of ows into
unemployment has exceeded 3 percent for several other sectors, including teaching (3.2
percent), construction (3.8 percent), transportation (3.7 percent) and production (3.4
percent). In the Great Recession, the only sector that exceeded 3 percent was construction.
These gures highlight how the high rate of labor market churn during the COVID-19
recession implies both a lower probability of being unemployed for two consecutive
months and a higher probability of becoming (newly) unemployed in any given month.


What are the most common transitions for workers formerly employed in the hardest-hit
Patterns of occupational transitions are mostly based on how well skills from one
occupation can transfer to another: Few fry cooks become plastic surgeons and vice versa.
However, the COVID-19 recession has seen increased interoccupational mobility. Among
the hardest-hit sectors, the percentage of workers continuously employed in those sectors
has been considerably lower during the COVID-19 recession than during the Great
Recession. In addition, the reemployment rate has been generally lower than during the
Great Recession — signaling again that this recession is characterized by more churn in the
labor market than other periods. For example, during the Great Recession, 89.4 percent of
workers employed in food occupations in any given month would still be employed in that
sector in the following month. In 2020, this percentage was 80.8 percent. Similarly, personal
care workers have a "staying" rate of 82.5 percent versus 88.4 percent in the Great
Recession. Comparable patterns hold for maintenance/cleaning, sales, o ce/admin and —
in a more modest fashion — for all other occupations.
Food, personal care, maintenance/cleaning and, to some extent, o ce/admin and sales are
all occupations whose skills are similar, as evidenced by the fact that workers in these
occupations regularly cycle across them. As one would expect, the pace of occupational
reallocation slowed down in the Great Recession. But in the COVID-19 recession, the pace
of interoccupational churn has exceeded not only the Great Recession rate, but that of the
2016–17 expansion.

The COVID-19 recession has featured an unusually high pace of worker ows between
employment and nonemployment and across major occupational categories. Sales, food
preparation/service, maintenance/cleaning and personal services are the occupational
groups that account for most of the ow both into and out of nonemployment. In addition,
workers in these occupations have experienced higher rates of repeated job loss and
occupational change during the COVID-19 recession than during the Great Recession.
Given these patterns, it appears desirable for economic policy to pay attention to alleviating
income instability from repeated displacements and occupational switches in addition to
lack of income caused by long-term unemployment. Higher rates of churn challenge
workers in both the short term and the long term. In the short term, workers going in and
out of employment see their income uctuate from month to month — which a ects their
ability to plan ahead and pay for everyday goods and services. In the medium to long term,
frequent nonemployment spells and occupational switches a ect a worker’s potential wage
growth by interrupting the accumulation of rm-speci c and sector-speci c human capital.
Workers in the hardest-hit occupations have historically seen lower wage-growth rates


than, for example, workers in the production sector. Frequent interruptions of their career
paths, in terms of both employer and occupational changes, may further depress their
prospects for sustained wage growth.
Claudia Macaluso is an economist in the Research Department of the Federal Reserve Bank
of Richmond. Many thanks to Alex Wolman for his encouragement and comments on an
earlier version of this brief. Rosemary Coskrey provided excellent research assistance.


Nonemployment is a broader term than unemployment. It includes workers who are out of
work and searching for jobs, those who are out of work and not searching and those who have
exited the labor force. The Richmond Fed publishes a monthly measure of nonemployment, the
Hornstein-Kudlyak-Lange Non-Employment Index.

Steven J. Davis, R. Jason Faberman and John C. Haltiwanger, "The Flow Approach to Labor
Markets: New Data Sources and Micro-Macro Links," Journal of Economic Perspectives, Summer
2006, vol. 20, no. 3, pp. 3–26. See also Steven J. Davis, R. Jason Faberman and John C.
Haltiwanger, "The Establishment-Level Behavior of Vacancies and Hiring," Quarterly Journal of
Economics, May 2013, vol. 128, no. 2, pp. 581–622.

Steven J. Davis and John C. Haltiwanger, "Labor Market Fluidity and Economic Performance,"

National Bureau of Economic Research Working Paper No. 20479, September 2014.

Claudia Macaluso, "Skill Remoteness and Post-Layo Labor Market Outcomes," Manuscript,
April 10, 2019.

Lisa B. Kahn, "The Long-Term Labor Market Consequences of Graduating from College in a Bad
Economy," Labour Economics, April 2010, vol. 17, no. 2, pp. 303–316.

Steven J. Davis and Till von Wachter, "Recessions and the Costs of Job Loss," Brookings Papers

on Economic Activity, Fall 2011, vol. 42, no. 2, pp. 1–72. See also Hannes Schwandt and Till M.
von Wachter, "Socioeconomic Decline and Death: Midlife Impacts of Graduating in a Recession,"
National Bureau of Economic Research Working Paper No. 26638, January 2020.

Ayşegül Şahin, Murat Tasci and Jin Yan, "Unemployment in the Time of COVID-19: A Flow-Based
Approach to Real-Time Unemployment Projections," National Bureau of Economic Research
Working Paper No. 28445, February 2021.

IPUMS-CPS, University of Minnesota.


For more on long-term unemployment during and after the Great Recession, see Andreas
Hornstein, Thomas A. Lubik and Jessie Romero, "Potential Causes and Implications of the Rise in
Long-Term Unemployment," Federal Reserve Bank of Richmond Economic Brief No. 11-09,
September 2011.


This article may be photocopied or reprinted in its entirety. Please credit the author, source,
and the Federal Reserve Bank of Richmond and include the italicized statement below.
Views expressed in this article are those of the author and not necessarily those of the Federal
Reserve Bank of Richmond or the Federal Reserve System.

Subscribe to Economic Brief
Receive an email noti cation when Economic Brief is posted online:
Email address


Contact Us
Karl Rhodes
(804) 697-8144

© 1997-2021 Federal Reserve Bank of Richmond