Recovery from the COVID-19 Recession Among Young Workers

Summary

NBER Working Paper No. 29307, Recovery from the COVID-19 Recession: Uneven Effects among Young Workers?, by Pinka Chatterji and Yue Li, dated September 2021 and revised December 2022. The paper uses January 2016-October 2022 Current Population Survey data to examine labor market recovery among 15-19-year-olds and 20-24-year-olds and to test the effects of states ending pandemic unemployment insurance programs. It reports that teenagers recovered fully while recovery among 20-24-year-olds was sluggish and incomplete into 2022. It finds that termination of pandemic UI programs increased work hours and full-time employment among 20-24-year-olds but not among 15-19-year-olds. The paper sets out its regression equations and closes with result tables and appendix figures on work hours by occupation and full- and part-time employment.

Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used

Full text

NBER WORKING PAPER SERIES

RECOVERY FROM THE COVID-19 RECESSION:
UNEVEN EFFECTS AMONG YOUNG WORKERS?

Pinka Chatterji
Yue Li

Working Paper 29307
http://www.nber.org/papers/w29307

NATIONAL BUREAU OF ECONOMIC RESEARCH
1050 Massachusetts Avenue
Cambridge, MA 02138
September 2021, Revised December 2022

Yue Li has received grants from the NSF and SSA via the NBER for some other projects. The
views expressed herein are those of the authors and do not necessarily reflect the views of the
National Bureau of Economic Research.

NBER working papers are circulated for discussion and comment purposes. They have not been
peer-reviewed or been subject to the review by the NBER Board of Directors that accompanies
official NBER publications.

© 2021 by Pinka Chatterji and Yue Li. All rights reserved. Short sections of text, not to exceed
two paragraphs, may be quoted without explicit permission provided that full credit, including ©
notice, is given to the source.
Recovery from the COVID-19 Recession: Uneven Effects among Young Workers?
Pinka Chatterji and Yue Li
NBER Working Paper No. 29307
September 2021, Revised December 2022
JEL No. I0,J0

ABSTRACT

In this paper, we examine the labor market recovery from the COVID-19 recession and test for
effects of termination of pandemic unemployment insurance programs among 15-24-year-olds.
We use data from the January 2016-October 2022 Current Population Survey. Using regression-
based methods, we show that while 15-19-year-olds experienced a brisk, full recovery in labor
market outcomes from the COVID-19 recession, the recovery was sluggish and incomplete
among 20-24-year-olds, with some work outcomes lagging below pre-pandemic norms well into
2022. Termination of pandemic UI programs led to increased work hours and full-time
employment among 20-24-year-olds but did not have these effects among 15-19-year-olds.

Pinka Chatterji
State University of New York at Albany
Economics Department
1400 Washington Avenue
Albany, NY 12222
and NBER
pchatterji@albany.edu

Yue Li
Department of Economics
SUNY Albany
1400 Washington Avenue
Albany, NY 12222
USA
yli49@albany.edu
1. Introduction

Economic downturns tend to harm young people’s labor market outcomes

disproportionately. Young workers have limited job tenure, education, and experience, and they

tend to work in industries and occupations that are vulnerable to the business cycle (Hoynes et

al., 2012). Research suggests that higher unemployment rates are associated with lower hiring

rates, but only for the least experienced workers (Forsythe, 2022). Moreover, there is evidence

of “scarring effects,” or lasting negative impacts of current economic downturns on young

workers’ future labor market outcomes (Glatt & Wunnava, 2018). Jobs obtained during

recessions tend to be lower-paying and of lower quality compared to those found during

expansions, and these effects have been found to be particularly damaging for the least

experienced workers (Oreopoulos et al., 2012).

The COVID-19 downturn brought additional challenges for young workers compared to

those of a typical recession, mainly because young workers are disproportionately represented in

industries that were severely affected by the pandemic, such as leisure and hospitality, and

wholesale and retail trade (Alon et al., 2020; Baker et al., 2020; Falk et al., 2021; Aaronson &

Alba, 2020; Kochhar & Barroso, 2020). As expected, the initial labor market impact of the

pandemic was largest for young workers (as well as for other disadvantaged groups), although in

some cases, the hardest-hit groups also recovered the fastest (Lee et al. 2021). Prior studies have

examined the effects of the pandemic on particular demographic groups (examples include Lee

et al., 2021; CBPP, 2022; Goda et al. 2021; Borjas & Cassidy, 2020; Alon et al., 2021), but these

studies have not focused on labor market recovery among young workers.

In this paper, we examine the labor market recovery from the COVID-19 recession of

two groups of young workers - teenagers (15-19-year-olds) and young adults (20-24-year-olds)

3
- and we test for effects of termination of pandemic unemployment insurance (UI) programs in

these two age groups. Exploring potential heterogeneity within young workers (teenagers vs.

young adults) is important because: (1) teenagers have weaker labor force attachment than young

adults and thus are less likely than young adults to be eligible for unemployment insurance (UI);

and (2) effects of work on human capital accumulation may be quite different for teenaged

workers, who are typically enrolled in school, vs. for young adult workers, many of whom have

completed their educations and are building full-time work experience.

We use data from the January 2016-October 2022 Current Population Survey (CPS)

(Flood et al., 2022). Using regression-based methods, we show that while 15-19-year-olds

experienced a brisk, full recovery in labor market outcomes from the COVID-19 recession, the

recovery was sluggish and incomplete among 20-24-year-olds, with some work outcomes

lagging below pre-pandemic norms well into 2022. Termination of pandemic UI programs led to

increased work hours and full-time employment among 20-24-year-olds but did not have these

effects among 15-19-year-olds.

2. Background

The COVID-19 recession was the deepest and also the shortest recession in recent history

(CBPP, 2022). This recession was fueled by both demand-side and supply-side shocks caused, to

a large extent, by the need to stop the spread of disease by shutting down the economy

(Gopinath, 2020; Handwerker et al., 2020). In response to the pandemic, the US Congress passed

the $2 trillion Coronavirus Aid, Relief, and Economic Security (CARES) Act on March 27,

2020. The new law created Pandemic Unemployment Assistance (PUA) which granted UI

eligibility to some groups of workers who previously had been ineligible (e.g. self-employed

workers, freelancers). This was particularly important for young workers. In addition, the

4
CARES Act provided Federal Pandemic Unemployment Compensation (FPUC), a weekly

additional payment of $600 per week to calculated state UI benefits between April 5, 2020, and

July 26, 2020; and provided Pandemic Emergency Unemployment Compensation (PEUC), an

additional 13 weeks of UI once state UI benefits (which typically last 26 weeks) had expired,

available until December 31, 2020 (NYS DOL 2021; Mishory & Settner, 2020).

The $600 additional UI benefits were extended to September 6, 2021, at a reduced

amount of $300 under the Lost Wage Assistance program, the Coronavirus Response and Relief

Supplemental Appropriations Act of 2021, and the American Rescue Plan Act of 2021 (BEA,

2021). Twenty-six states, however, chose to end their participation in either or both FPUC and

PUA before September 2021. Eighteen of these twenty-six states ended both FPUC and PUA in

June 2021, while twenty-four states and DC continued both programs until the programs expired

in September 2021 (Holzer et al. 2021). The rationale for these early terminations of pandemic

UI was that that generous benefits possibly were dampening workers’ efforts to find jobs and

return to work (Holzer et al., 2021). Empirical research using a variety of data sets and study

designs indicates that generous UI benefits during the pandemic had limited or no effects on

employment (Altonji et al., 2020; Bartik et al., 2020; Finamor & Scott, 2021; Petrosky-Nadeau

& Valletta, 2021; Ganong et al., 2021a-b). 1 Researchers have started to estimate the effect of

some states’ early termination of pandemic FPUC and PUA in June and July 2021 on

employment, and this work shows mixed effects (Holzer et al., 2021; Dube, 2021; Coombs et al.

2021).

All young workers have relatively low levels of job tenure, skills, and experience, but the

15-19 vs. 20-24 age groups differ in two respects that are relevant to the time-period following

1
Marinescu et al. (2021) find that FPUC was associated with a 3.6 percent decrease in job applications
and hence increased labor market tightness, which was depressed under the FPUC period.

5
the COVID-19 recession. First, workers aged 15-19 face higher costs of working than those aged

20-24 since the vast majority of teenagers are enrolled in school; thus, a lower share of this age

group is employed (as we discuss below). Second, relative to young adults, teenagers are less

likely to be eligible for unemployment insurance (UI) because of failing to meet state regulations

such as having sufficient work history (Mishory & Stetter, 2020). For these reasons, we

hypothesize that there was heterogeneity among young workers in labor market recovery from

the COVID-19 recession.

This heterogeneity within young workers is important to study from a policy perspective.

From a short-term policy perspective, firms’ hiring of less-experienced teen workers during the

pandemic may have benefitted both teenage workers as well as young adult workers who

reduced work hours and received enhanced UI. The long-term policy implications, however, are

unclear since young adult workers’ later labor market outcomes may be harmed by these early

disruptions in work (Glatt & Wunnava, 2018; Santacrose, 2013). Moreover, the long-term effect

on the younger group of teenaged workers is ambiguous, since working during the school year

could negatively impact academic performance and time spent studying (Tyler, 2003; Rothstein,

2007; Kalenkoski & Pabilonia, 2012) while working during summer months may improve

academic performance (Leos-Urbel, 2014).

3. Data and Methods

The analysis sample includes childless respondents aged 15-24 who do not have college

degrees; do not reside in group quarters; and are US citizens who were in a CPS household for at

least one month between January 2016 and October 2022. Initially, using this sample, we

estimate Equation 1 to map out the month-by-month effects of the pandemic on work outcomes

for each age group (15-19, 20-24) during the pandemic:

6
𝜏𝜏 𝜏𝜏 32 𝜏𝜏 𝜏𝜏
𝑦𝑦𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 = ∑−1
𝜏𝜏=−13 𝛿𝛿 𝐷𝐷𝑖𝑖𝑖𝑖 + ∑𝜏𝜏=1 𝛿𝛿 𝐷𝐷𝑖𝑖𝑖𝑖 + 𝜂𝜂𝑡𝑡𝑡𝑡 + 𝛾𝛾𝑎𝑎𝑎𝑎 + 𝛾𝛾𝑠𝑠 + 𝑋𝑋𝑖𝑖𝑎𝑎𝑎𝑎𝑎𝑎 𝜃𝜃 + 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 (1)

In Equation 1, 𝑦𝑦𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 is a labor market outcome (described below) for individual 𝑖𝑖 of age a

living in state s at time 𝑡𝑡. The term 𝐷𝐷𝑖𝑖𝑖𝑖𝜏𝜏 represents a set of indicators for the 13 months before the

𝜏𝜏 𝜏𝜏
pandemic (Jan 2019 – Jan 2020; ∑−1
𝜏𝜏=−13 𝛿𝛿 𝐷𝐷𝑖𝑖𝑖𝑖 ) and 32 months during the pandemic (Mar 2020-

𝜏𝜏 𝜏𝜏
Oct 2022; ∑32
𝜏𝜏=1 𝛿𝛿 𝐷𝐷𝑖𝑖𝑖𝑖 ), with Feb 2020 normalized to 0. Equation 1 also includes the full set of

interactions between each of the 12 calendar months and each of the 5 age dummies (𝛾𝛾𝑎𝑎𝑎𝑎 ) to

control for differential seasonal patterns by age; state fixed effects (𝛾𝛾𝑠𝑠 ); and a linear monthly

time trend in time (𝑡𝑡𝑡𝑡 ) which is set to be unchanged after Feb 2020 to capture the slowdown of

economic activity during the pandemic.

The individual controls (𝑋𝑋𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 ) include dummy variables for: female, Black, other

race/ethnicity, Hispanic origin, and metropolitan status (in central city and outside central city,

with other as the baseline). The estimated coefficients 𝛿𝛿 1 − 𝛿𝛿 32 capture the month-by-month

effects of the pandemic — the difference in outcomes in each month during the pandemic

relative to the corresponding month in prior years. For all models, we apply CPS person weights,

and estimate robust standard errors with two-way clustering on state and year-month.

Next, we explore potential heterogeneity within young workers in (1) the year-by-year

effects of the pandemic; and (2) the effects of early termination of pandemic UI programs on

labor market outcomes. To do so, for each age group (15-19 and 20-24), we estimate Equation 2,

in which we replace the month indicators with year indicators for 2020, 2021, and 2022. We also

estimate Equation 3, which adds to the Equation 2 model a dummy variable that is set equal to

one starting in the month after pandemic UI had been terminated in this state (zero otherwise).

𝑦𝑦𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 = 𝛽𝛽1 𝑌𝑌𝑌𝑌𝑌𝑌𝑌𝑌2020𝑖𝑖𝑖𝑖 + 𝛽𝛽2 𝑌𝑌𝑌𝑌𝑌𝑌𝑌𝑌2021𝑖𝑖𝑖𝑖 + 𝛽𝛽3 𝑌𝑌𝑌𝑌𝑌𝑌𝑌𝑌2022𝑖𝑖𝑖𝑖 + 𝜂𝜂𝑡𝑡𝑡𝑡 + 𝛾𝛾𝑎𝑎𝑎𝑎 + 𝛾𝛾𝑠𝑠 + 𝑋𝑋𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝜃𝜃 + 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖

(2)

7
𝑦𝑦𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 = 𝛽𝛽1 𝑌𝑌𝑌𝑌𝑌𝑌𝑌𝑌2020𝑖𝑖𝑖𝑖 + 𝛽𝛽2 𝑌𝑌𝑌𝑌𝑌𝑌𝑌𝑌2021𝑖𝑖𝑖𝑖 + 𝛽𝛽3 𝑌𝑌𝑌𝑌𝑌𝑌𝑌𝑌2022𝑖𝑖𝑖𝑖 + 𝛽𝛽4 𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑛𝑛𝑖𝑖𝑖𝑖𝑖𝑖 + 𝜂𝜂𝑡𝑡𝑡𝑡 + 𝛾𝛾𝑎𝑎𝑎𝑎 +

𝛾𝛾𝑠𝑠 + 𝑋𝑋𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝜃𝜃 + 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 (3)

In Equations 2-3, Year 2020 is a 0-1 indicator for Mar 2020-Dec 2020; Year 2021 is a 0-1

indicator for the entire year of 2021; and Year 2022 is a 0-1 indicator for Jan 2022-October 2022.

The UI expiration variable is a 0-1 indicator for the months after which state 𝑠𝑠 had terminated at

least one enhanced UI program (FPUC or PUA). 2 We conduct several robustness checks of our

findings using Equation 3; results are discussed below.

As outcomes in our analysis, we consider usual work hours (all occupations), usual work

hours by occupation, employed, employed full-time and employed part-time. The variable “usual

work hours” captures hours usually worked per week at all jobs, or the number of hours worked

last week if the respondent reports having a flexible work schedule (including zeros). We

examine usual work hours in all occupations and usual work hours by the following three

occupational categories: (1) food preparation and serving; (2) sales and related occupations; and

(3) all other occupations. These occupational categories were selected because they are the

largest occupational categories for young workers. 3 The variable “employed” is an indicator for

2
The termination month is set to June 2021 for Alabama, Alaska, Arkansas, Florida, Georgia, Idaho,
Iowa, Mississippi, Missouri, Montana, Nebraska, New Hampshire, North Dakota, Ohio, Oklahoma, South
Carolina, South Dakota, Texas, Utah, West Virginia and Wyoming, and July 2021 for Tennessee, Arizona
and Louisiana. As Indiana and Maryland are ordered by a court to resume participation in enhanced UI
programs, we set the termination month to be September 2021 for these two states and the remaining 24
states and DC (Congressional Research Service 2021).
3
As of 2019, among employed 15-24-year-olds in the CPS, the two most common occupations were
“food preparation and serving” and “sales and related.” Among 15-19-year-old workers in 2019, 27 and
21 percent reported food and sales occupations, respectively. Among 20-24-year-old workers in 2019, 20
and 17 percent reported food and sales occupations, respectively. Occupation is characterized based on
the Census Bureau’s 2010 occupation classification. A respondent is classified working in food
preparation and serving occupations if his/her occupation code is between 4000-4150, and sales related
occupations if the code is between 4700-4965, and other occupations if the code is outside of the
abovementioned ranges.

8
whether a respondent is at work in the preceding week. The variables “works full-time” and

“works part-time” are based on the respondents’ employment statuses.

In Appendix Table 1, we show summary statistics for the 15-19-year-old and the 20-24-

year-old samples. In the 15-19-year-old sample, 49 percent of respondents are female, 15 percent

are Black, 23 percent are Hispanic, and 11 percent are classified as having other race/ethnicity.

In the 15-19-year-old age group, 25 percent live in a central city, and 48 percent live outside a

central city. These demographic characteristics are similar to those of the 20-24-year-old sample

with a few exceptions. First, only 45 percent of the 20-24-year-old sample is female, probably

because the samples are restricted to childless respondents without college education. Also, 20-

24-year-olds are more likely than 15-19-year-olds to be living in central cities.

As one would expect, 15-19-year-olds and 20-24-year-olds differ in their labor force

attachment. Teenagers are much less likely than young adults to be employed full-time (6

percent for 15-19-year-olds vs. 36 percent for 20-24-year-olds), and they are much more likely

than young adults to be currently enrolled in school (60 percent for 15-19-year-olds vs. 41

percent for 20-24-year-olds) (Appendix Table 1). In Appendix Table 2, we show employment

rates, school enrollment rates, and employment rates in the month of July (when school is not in

session) conditional on being enrolled in school in the month of April (when school is in

session). The 15-19-year-old respondents’ employment is, to some extent, more seasonal that

that of 20-24-year-old respondents (although both age groups’ outcomes show strong

seasonality). For both age groups, full-time employment status is substantially higher in July vs.

in April, but part-time employment status increases in July vs. April only for the 15-19-year-old

sample. These seasonal effects are driven by respondents who are in school as of April

(Appendix Table 2).

9
4. Results

Figures 1 and 2 show the estimated coefficients on the month indicators in Equation 1

(𝛿𝛿 𝜏𝜏 ). Figure 1 shows estimates from the model of work hours (in all occupations), while Figure 2

shows estimates from the employment model. In Appendix Figures 1-5, we show these figures

for the remaining outcomes – work hours in food and serving occupations, work hours in sales

and related occupations, work hours in other occupations, employed full-time, and employed

part-time. As a group, these figures highlight the impact of the pandemic on work outcomes

among 15-19-year-old and 20-24-year-old CPS respondents from January 2019 to October 2022.

Figure 1 shows that work hours (all occupations), which were stable in 2019, plunged by

about 1.75 hours among 15-19-year-olds and by 6.89 hours among 20-24-year-olds in April

2020; these sudden drops represent about 30 and 32 percent decreases, respectively, evaluated at

the pre-pandemic means (5.94 hours and 21.49 hours, Appendix Table 1). Work hours among

15-19-year-olds had returned to pre-pandemic norms by September 2020. In fact, by April 2021,

work hours in the 15-19-year-old age group had started to exceed typical levels, often by about

10 percent, evaluated at the pre-pandemic mean. As of October 2022, work hours among 15-19-

year-olds were still 10 percent above their typical level.

Among 20-24-year-olds, although work hours (all occupations) steadily recovered from

April 2020 until November 2020, progress started to lag at that point. Even in August 2021,

work hours were still 2.1 hours below normal levels, which is about 10 percent lower than the

pre-pandemic mean for the 20-24-year-old age group. We note that for 20-24-year-olds there

appears to be a sharp, short-lived bounce back of overall work hours starting in September 2021,

which coincides with the expiration of pandemic UI benefits in all states; we do not observe this

10
bounce back among 15-19-year-olds. As of October 2022, work hours among 20-24-year-olds

were at their typical pre-pandemic levels.

Figure 2 shows the same estimates, but for the model of employment. Along this margin,

we see similar effects – as of April 2020, the pandemic had reduced employment by 11

percentage points (about 44 percent relative to the pre-pandemic mean) among 15-19-year-olds,

and by 24 percentage points (about 38 percent relative to the pre-pandemic mean) among 20-24-

year-olds. In the younger age group, employment had recovered by September 2020, but in the

older group, negative effects persisted until a temporary bounce back in October 2021. By

January 2022, employment was once again persistently below pre-pandemic levels among 20-24-

year-olds. The persistent, negative effects of the pandemic on work outcomes among 20-24-year-

olds appear to be driven by reductions in part-time employment (Appendix Figures 5). By

October 2022, employment was back at pre-pandemic levels for 20-24-year-olds but was still

about 8 percent higher among 15-19-year-olds relative to the pre-pandemic mean.

Table 1 shows estimates from Equation 2, in which we estimate the yearly effects of the

pandemic on work hours and employment by age group (15-19, Panel A vs. 20-24, Panel B). The

findings show that the pandemic had negative effects on work outcomes in both age groups, but

among 15-19-year-olds, the negative effects occurred in 2020 only. Among 15-19-year-olds, the

pandemic is associated with an 11 percent reduction in work hours (all occupations), a 17 percent

reduction in full-time employment, and a 16 percent reduction in part-time employment in 2020;

for the 20-24-year-old age group, these magnitudes were 17 percent, 22 percent, and 12 percent

respectively (relative to the pre-pandemic means, Appendix Table 1).

By 2021, however, outcomes in the 15-19-year-old group had more than fully recovered,

with work hours and employment levels persisting at higher-than-typical levels into 2022. For

11
example, in 2021 and 2022, work hours (all occupations) were still 6-8 percent higher-than-

typical for 15-19-year-olds (Panel A, Table 1). In contrast, in the 20-24-year-old age group, the

pandemic’s toll on work outcomes continued into 2021, with some outcomes still showing

adverse effects in 2022. For example, as of 2022 among 20-24-year-olds, the pandemic is

associated with a 17 percent reduction in work hours in food and serving occupations, a 3

percent decrease in any employment, and a 4 percent reduction in part-time employment.

This raises the question of pandemic UI, and whether teenagers returned to work more

quickly than young adults because they were less likely to be eligible for benefits than young

adults. From the findings in Table 2, we gain insight into this possibility by testing whether the

expiration of pandemic UI benefits, which varied by state, had differing effects by age group. In

Panel A, we show estimates of Equation 3 generated using the 15-19-year-old sample. These

findings show that the expiration of pandemic UI programs has no statistically significant effects

on outcomes except for work hours in sales occupations; expiration of pandemic UI programs is

associated with (counter-intuitively) a 15 percent decrease in work hours in sales occupations,

relative to pre-pandemic means. For the older age group (20-24), expiration of pandemic UI

programs is associated with about a 6 percent increase in work hours (all occupations) and a 6

percent increase in full-time employment, relative to pre-pandemic means. These effects are

driven by increases in work hours in the “other” occupation category.

In Table 3, we show findings from several robustness checks, based on the model shown

in Table 2 (Equation 3). First, we consider the possible effects of pandemic-era CPS nonresponse

bias on our findings. Ward & Edwards (2021) provide evidence that the disruption and cessation

of in-person interviewing during the pandemic affected the composition of the CPS sample in

ways that the sampling weights do not adequately address. Specifically, it appears that the abrupt

12
change in interviewing mode led to a shift in the CPS sample towards more advantaged

individuals with greater-than-average attachment to the labor force. For our analysis, this finding

may suggest that we are underestimating the negative effects of the pandemic on outcomes, if the

most disadvantaged young workers have dropped out of the CPS sample.

To address this point, we estimate the work hours (all occupations) model limiting the

samples to CPS respondents who could not be reached for their first or fifth Months-in-sample

(MIS) interview, i.e., those who skipped MIS 1 interview but only responded between MIS 2-4

and those who skipped MIS 5 and only responded between MIS 6-8 (referred to as MIS2+). As

discussed in Ward & Edwards (2021), the MIS2+ group includes harder-to-reach and more

disadvantaged individuals; thus, we may expect more harmful effects of the pandemic on their

outcomes and a less robust rebound compared to estimates based on our main sample. In Table 3,

columns 1-2, however, we see that the estimates for the MIS2+ group are similar to our main

findings, with two exceptions. First, we no longer see the significantly increased work hours in

2021 and 2022 in the 15-19-year-old sample when we limit this sample to the MIS2+ sub-

sample. Second, expiration of pandemic UI programs is not associated with a significant increase

in work hours (all occupations) in the MIS2+ sample.

Next, as a robustness check, we include controls for monthly COVID-19 cases and

deaths for each state as covariates in the models. At a 10% significance level, COVID-19 deaths

are negatively related to work hours among 15-19-year-olds, perhaps due to parents’ fears about

teenagers’ exposure to the disease in the work setting. Including these covariates only sharpens

our conclusions from the main findings; we see persistent negative effects of the pandemic on

20-24-year-olds; an improvement in work hours later in the pandemic for 15-19-year-olds; and

13
an increase in work hours associated with early termination of pandemic UI programs for 20-24-

year-olds.

Finally, we explore whether the way in which we measured respondents’ ages affected

our results. In the main analyses, we limit the samples to people aged 15-19 or 20-24 at the time

of the CPS interview. Alternatively, the age group could be determined based on the

respondent’s age at the start time of the pandemic, and then respondents could be followed

afterward. We try this alternative method, re-creating the samples by limiting them to

respondents who were 15-19 or 20-24 years old in 2020. 4 After setting up the samples in this

way, we re-estimated the “work hours (all occupations)” model shown in Table 2. This

robustness check is shown in columns 5-8 of Table 3. We show specifications with and without

the age × month interactions because of the limited variation within age by month cells. Overall,

the magnitudes of the estimates change somewhat, but the qualitative pattern of results is similar

to that of our main findings.

Finally, in Appendix Table 3, we re-estimated the “usual work hours (all occupations)”

model based on Equation 3 for three additional age groups – ages 25-54, 55-64, and 65-74.

Columns 1-3 show findings for these age groups when samples are limited to childless

respondents without college education, while columns 4-6 show estimates based on the samples

for these age groups without such restrictions. We note that the pattern of findings for 25-54-

year-olds looks similar to that of the 20-24-year-old group in our study – the pandemic has

persistent effects on work hours, and pandemic UI programs appear to have played a role.

4
Because our samples are also limited to childless respondents without 4-year college degrees, the
compositions of the samples can still change, even though we are now following the same cohorts over
time. Also, this sample restriction leads to some of the younger cohorts having limited or no pre-periods;
for example, if a respondent is aged 15 in 2020, we only observe this person’s outcomes post-pandemic.
In addition, in some cases, we lack variation in age by month cells – for example, we only observe people
aged 26 in a single year (2022).

14
Pandemic UI programs generally are not associated with work hours (all occupations) among

workers aged 55 and older.

5. Discussion and Conclusions

Our findings show heterogeneity by age group in the effects of the pandemic among

young, childless workers without college degrees. The youngest workers, aged 15-19, were more

resilient than their 20-24-year-old counterparts. There was a brisk, full recovery in labor market

outcomes among 15-19-year-olds. Among 20-24-year-olds, however, the recovery was sluggish,

with work hours lagging below pre-pandemic norms, well into the year 2022.

Our findings support the idea that the heterogeneity by age group may be due in part to

enhanced UI benefits, which were more accessible to 20-24-year-olds vs. 15-19-year-olds.

Enhanced UI benefits may have made it possible for 20-24-year-old workers to turn down

employment offers, perhaps to hold out for more favorable terms, or invest in education and

training (Levitz, 2021), contributing to the tightness of the labor market. The heterogeneity may

stem from other reasons as well. Fear of contagion may be a more pressing issue for 20-24-year-

old workers, because they work more hours than 15-19-year-olds and are more likely to be living

with non-relatives, and as a result have greater exposure to COVID-19. In addition, 20-24-year-

old workers have more experience than 15-19-year-old workers and thus may have jobs that

involve more person-to-person contact and greater COVID-19 exposure risk.

Fear of contagion is likely to have been an important factor in the initial waves of the

pandemic, when vaccinations were not widely available, and disease was relatively severe. With

widespread vaccination, fear of contagion may have been less of a concern for workers by the

second half of 2021. The expiration of enhanced UI benefits in all states in September 2021,

15
therefore, may have induced 20-24-year-old workers to return to their pre-pandemic work hours

in this month; our findings are consistent with this story.

Our study has important limitations. First, non-response bias during the pandemic may

have affected our findings in ways that our simple robustness check does not capture. We may be

under-stating the negative effects of the pandemic and over-stating the recovery if the most

disadvantaged workers left the CPS sample during the pandemic, and if the sampling design does

not adequately address this issue. Second, states and localities may have implemented other

policies that coincided with the termination of pandemic UI programs, and these unmeasured

factors may have affected our estimates of the effects of termination on work outcomes.

As of December 2022, COVID-19 continues to chart an unpredictable course throughout

the world. In this challenging environment, it is critical to inform policymakers by estimating the

impact of the pandemic on young workers, and by exploring heterogenous effects within this

group. Future research should address these longer-term effects of the pandemic on young

workers’ human capital and labor market outcomes.

16
References

1. Aaronson, S. & Alba, F. (April 15, 2020). “The unemployment impacts of COVID-19:

lessons from the Great Recession.” Available at: The unemployment impacts of COVID-

19: lessons from the Great Recession (brookings.edu). Accessed 9/9/21.

2. Alon, T.M., Doepke, M., Rumsey-Olmstead, J., & Tertilt, M. (April 2020). The impact of

COVID-19 on gender equality. NBER working paper 26947. Available at:

http://www.nber.org/papers/w26947.

3. Alon, T.M., Coskun, S., Doepke, M., Koll, D. & Tertlit, M. (2021). “From mancession to

shecession: Women’s employment in regular and pandemic recessions.” NBER

Macroeconomics Annual, 36 (Eds. Eichenbaum & Hurst).

4. Altonji, J., Contractor, Z., Finamor, L., Haygood, R., Lindenlaub, I., Meghir, C., O’Dea,

C., Scott, D., Wang, L., & Washington, E. (July 14, 2020). “Employment effects of

unemployment insurance generosity during the pandemic.” Available at:

https://tobin.yale.edu/sites/default/files/files/C-19%20Articles/CARES-

UI_identification_vF%281%29.pdf, Accessed 5/22.

5. Baker, S.R., Farroknhia, R.A., Meyer, S., Pagel, M., & Yannelis, C. (July 21 2020).

“How does household spending respond to an epidemic? Consumption during the 2020

COVID-19 pandemic.” Rev Asset Pricing Stud..

6. Bartik, A.W., Bertrand, M., Lin, F., Rothstein, J. & Unrath, M. (2020). "Measuring the

Labor Market at the Onset of the COVID-19 Crisis," Brookings Papers on Economic

Activity, vol 2020(2), pages 239-268.

17
7. Borjas, G.J. & Cassidy, H. (2020). “The adverse effect of the COVID-19 labor market

shock on immigrant employment.” NBER working paper 27243.

8. Bureau of Economic Analysis (BEA), “How will the expansion of unemployment

benefits in response to the COVID-19 pandemic be recorded in the NIPAS?” Available

at: https://www.bea.gov/help/faq/1415, Accessed 9/10/21.

9. Center on Budget and Policy Priorities (CBPP), Chart book: Tracking the recovery from

the pandemic recession. Available at: https://www.cbpp.org/research/economy/tracking-

the-recovery-from-the-pandemic-recession. Accessed 12/22.

10. Congressional Research Service (2021) States Opting Out of COVID-19 Unemployment

Insurance (UI) Agreements IN11679. Last updated August 20, 2021.

11. Coombs, K., Dube, A., Jahnke, C., Kluender, R., Naidu, S., & Stepner, M. (August 20,

2021). “Early withdrawal of pandemic unemployment insurance: Effects on earnings,

employment, and consumption.” Working paper.

12. Dube, Arindrajit. (August 20, 2021). “Early withdrawal of pandemic UI: impact on job

finding in July using Current Population Survey.” Available at:

https://arindube.com/blog-posts/. Accessed 9/9/21.

13. Falk, G., Romero, P.D., Carter, J.A., Nicchitta, I.A., & Nyhof, E.C. (June 15, 2021).

“Unemployment rates during the COVID-19 pandemic,” Congressional Research Service

Report 46554, Available at: https://crsreports.congress.gov/. Accessed 9/21.

14. Finamor, L. & Scott, D. “Labor market trends and unemployment insurance generosity

during the pandemic.” Economics Letters 199.

18
15. Flood, S. King, M., Rodgers, R., Ruggles, R., Warren, J.R. and Westberry, M. (2022)

Integrated Public Use Microdata Series, Current Population Survey: Version 10.0

[dataset]. Minneapolis, MN: IPUMS.

16. Forsythe, E. (2021). “Why don’t firms hire young workers during recessions?” The

Economic Journal 132: 1765-1789.

17. Ganong, P., Greig, F., Liebeskind, M., Noel, P. Sullivan, D.M. & and Vavra, J. (February

2021a). “Spending and Job Search Impacts of Expanded Unemployment Benefits:

Evidence from Administrative Micro Data.” Becker Friedman Institute for Economics

Working Paper 2021-19.

18. Ganong, P., Greig, F., Noel, P., Sullivan, D.M. & Vavra, J. (July 29, 2021b). “Micro and

macro disincentive effects of expanded unemployment benefits.” Working paper.

19. Glatt, J. & Wunnava, P. (2018). “Help Not Wanted: The Dismal Science of Youth

Unemployment’s Scarring Effect.” iBusiness, 10, 51-84. Available at:

http://www.scirp.org/journal/ib.

20. Goda Shah, G., Jackson, E., Nicholas, L.H. & Stith, S.S. (2021). “The impact of COVID-

19 on older workers’ employment and social security spillovers.” NBER working paper

29083.

21. Gopinath, G. (April 2020). Limiting the economic fallout of the coronavirus with large

targeted policies in Mitigating the COVID-19 Crisis – Act Fast and Do Whatever it

Takes.

22. Handwerker, E.W., Meyer, P.B. & Piacentini, J. (December 2020). “Employment

recovery in the wake of the COVID-19 recession.” US Bureau of Labor Statistics

19
Monthly Labor Review, Available at:

https://www.bls.gov/opub/mlr/2020/article/employment-recovery.htm; Accessed 9/2021.

23. Holzer, H.J., Hubbard, R.G. & Strain, M.R. (December 2021). “Did pandemic

unemployment benefits reduce employment? Evidence from early state-level expirations

in June 2021.” NBER Working Paper 29575.

24. Hoynes, H., Miller, D.L. & Schaller, J. (Summer 2012). “Who suffers during recessions?”

Journal of Economic Perspectives 26:3, 27-48.

25. Kalenkoski, C.M. & Pabilonia, S. W. (2012). “Time to work or time to play: The effect

of student employment on homework, sleep, and screen time.” Labour Economics, 19,

211-221.

26. Kochhar, R. & Barroso, A. (March 20, 2020). “Young workers likely to be hard hit as

COVID-19 strikes a blow to restaurants and other service sector jobs.” Pew Research

Center, Available at: https://www.pewresearch.org/fact-tank/2020/03/27/young-workers-

likely-to-be-hard-hit-as-COVID-19 -strikes-a-blow-to-restaurants-and-other-service-

sector-jobs/. Accessed 9/9/21.

27. Lee, S.Y., Park, M. & Shin, Y. (2021). “Hit harder, recover slower? Unequal

employment effects of the COVID-19 shock.” NBER working paper 28354.

28. Leos-Urbel, J. (2014). “What Is a Summer Job Worth? The Impact of Summer Youth

Employment on Academic Outcomes.” Journal of Policy Analysis and Management Vol.

33, No. 4, pp. 891-911

29. Levitz, E. (May 7, 2021). “5 Explanations for April’s Bad Jobs Report.” Available at:

https://nymag.com/intelligencer/2021/05/jobs-report-explained-ui-childcare-seasonal-

adjustment.html. Accessed: 9/10/21.

20
30. Marinescu, I., Skandalis, D. & Zhao, D. (2021). "The impact of the Federal Pandemic

Unemployment Compensation on job search and vacancy creation," Journal of Public

Economics, Elsevier, vol. 200(C).

31. Mishory, J. & Stettner, A. (April 23, 2020). “Unemployment insurance and young people

in the wake of COVID-19.” The Century Foundation. Available at: Unemployment

Insurance and Young People in the Wake of COVID-19 (tcf.org); Accessed 9/21.

32. New York State Department of Labor (NYS DOL), “Coronavirus aid, relief, and

economoc security (CARES) act.” Available at: https://dol.ny.gov/coronavirus-aid-relief-

and-economic-security-cares-act. Accessed 9/21.

33. Oreopoulos, P., von Wachter, T. & Heisz, A. (2012). “The short- and long-term career

effects of graduating in a recession.” AEJ: Applied Economics Vol. 4 No. 1: 1-29.

34. Petrosky-Nadeau, Nicolas, Robert G. Valletta. 2021 “UI Generosity and Job Acceptance:

Effects of the 2020 CARES Act,” Federal Reserve Bank of San Francisco Working Paper

2021-13. https://doi.org/10.24148/wp2021-13.

35. Rothstein, D.S. (2007). “High School Employment and Youths' Academic

Achievement.” The Journal of Human Resources, Vol. 42, No. 1 pp. 194-213.

36. Santarosa, R. (August 21, 2013). “Long-term consequences of youth unemployment.” US

Chamber of Commerce Foundation. Available at:

https://www.uschamberfoundation.org/blog/post/long-term-consequences-youth-

unemployment/34032. Accessed 5/22.

37. Tyler, J.H. (2003). “Using State Child Labor Laws to Identify the Effect of School‐Year

Work on High School Achievement.” Journal of Labor Economics, Vol. 21, No. 2, pp.

381-408.

21
38. Ward, J.M. & Edwards, K.A. (2021). “CPS nonresponse during the COVID-19

pandemic: Explanations, extent and effect.” Labour Economics, Vol. 72, 1020-60.

22
Table 1: Effect of pandemic on labor market outcomes

Panel A: 15-19 age group
Usual Work Hours Employment
All occ. Food Sales Other Any Full-time Part-time
(1) (2) (3) (4) (5) (6) (7)
Year 2020 -0.65*** -0.22** -0.11 -0.32*** -0.04*** -0.01*** -0.03***
(0.22) (0.08) (0.07) (0.12) (0.01) (0.00) (0.01)
Year 2021 0.38** 0.18** -0.03 0.23** 0.01 0.00* 0.00
(0.16) (0.08) (0.06) (0.11) (0.01) (0.00) (0.00)
Year 2022 0.47*** 0.17*** 0.00 0.30*** 0.02*** 0.00 0.01***
(0.15) (0.06) (0.05) (0.10) (0.00) (0.00) (0.00)
Y2020=Y2021 0.00 0.00 0.21 0.00 0.00 0.00 0.00
Y2020=Y2022 0.00 0.00 0.18 0.00 0.00 0.00 0.00
Y2021=Y2022 0.49 0.82 0.62 0.49 0.07 0.54 0.03
Panel B: 20-24 age group
Usual Work Hours Employment
All occ. Food Sales Other Any Full-time Part-time
(1) (2) (3) (4) (5) (6) (7)
Year 2020 -3.68*** -1.00*** -0.29** -2.38*** -0.11*** -0.08*** -0.03***
(0.56) (0.13) (0.13) (0.40) (0.02) (0.01) (0.01)
Year 2021 -1.09*** -0.54*** -0.08 -0.48 -0.04*** -0.02*** -0.02***
(0.34) (0.11) (0.11) (0.37) (0.01) (0.01) (0.01)
Year 2022 -0.40 -0.52*** -0.17 0.29 -0.02** -0.01 -0.01***
(0.33) (0.12) (0.16) (0.40) (0.01) (0.01) (0.00)
Y2020=Y2021 0.00 0.00 0.11 0.00 0.00 0.00 0.19
Y2020=Y2022 0.00 0.00 0.45 0.00 0.00 0.00 0.11
Y2021=Y2022 0.02 0.88 0.44 0.01 0.00 0.04 0.45

Note: *** p<0.01, ** p<0.05, * p<0.1. Table presents estimates of 𝛽𝛽s from equation (2). The number of
observations is 560,397 for Panel A and 360,175 for Panel B. Year 2020 is a 0-1 indicator for Mar 2020-
Dec 2020. Year 2021 is a 0-1 indicator for the entire year of 2021. Year 2022 is a 0-1 indicator for Jan
2022-Oct 2022. The last three rows report the T-test for equality of the estimated coefficients for two
years.

23
Table 2: Effect of terminating enhanced UI benefits

Panel A: 15-19 age group
Usual Work Hour Employment
All occ. Food Sales Other Any Full-time Part-time
(1) (2) (3) (4) (5) (6) (7)
Year 2020 -0.65*** -0.22** -0.11 -0.32*** -0.04*** -0.01*** -0.03***
(0.21) (0.08) (0.07) (0.12) (0.01) (0.00) (0.01)
Year 2021 0.46** 0.23** 0.03 0.20* 0.01 0.01** 0.00
(0.17) (0.10) (0.06) (0.11) (0.01) (0.00) (0.01)
Year 2022 0.69*** 0.30** 0.18** 0.20 0.02** 0.01** 0.02*
(0.19) (0.11) (0.07) (0.16) (0.01) (0.00) (0.01)
UI Expiration -0.21 -0.13 -0.18** 0.09 -0.01 -0.00 -0.00
(0.16) (0.08) (0.07) (0.16) (0.01) (0.00) (0.01)
Y2020=Y2021 0.00 0.00 0.03 0.00 0.00 0.00 0.00
Y2020=Y2022 0.00 0.00 0.00 0.01 0.00 0.00 0.00
Y2021=Y2022 0.06 0.18 0.02 0.96 0.04 0.86 0.02
Panel B: 20-24 age group
Usual Work Hour Employment
All occ. Food Sales Other Any Full-time Part-time
(1) (2) (3) (4) (5) (6) (7)
Year 2020 -3.68*** -1.00*** -0.29** -2.38*** -0.11*** -0.08*** -0.03***
(0.57) (0.13) (0.13) (0.41) (0.02) (0.01) (0.01)
Year 2021 -1.55*** -0.63*** -0.06 -0.86** -0.05*** -0.03*** -0.02***
(0.33) (0.13) (0.12) (0.34) (0.01) (0.01) (0.01)
Year 2022 -1.70*** -0.78*** -0.13 -0.79* -0.05*** -0.03** -0.01
(0.54) (0.23) (0.23) (0.46) (0.02) (0.01) (0.01)
UI Expiration 1.25** 0.25 -0.04 1.04** 0.02 0.02** 0.00
(0.50) (0.22) (0.16) (0.40) (0.01) (0.01) (0.01)
Y2020=Y2021 0.00 0.02 0.13 0.00 0.00 0.00 0.25
Y2020=Y2022 0.01 0.35 0.51 0.01 0.01 0.00 0.29
Y2021=Y2022 0.68 0.35 0.68 0.83 0.62 0.90 0.63
Note: *** p<0.01, ** p<0.05, * p<0.1. Table presents estimates of 𝛽𝛽s from equation (3). The number of
observations is 560,397 for Panel A and 360,175 for Panel B. Year 2020 is a 0-1 indicator for Mar 2020-
Dec 2020. Year 2021 is a 0-1 indicator for the entire year of 2021. Year 2022 is a 0-1 indicator for Jan
2022-Oct 2022. The last three rows report the T-test for equality of the estimated coefficients for two
years.

24
Table 3: Robustness checks

Usual Work Hour
MIS 2+ sample + Covid controls Define age group using age in 2020
w/o age×month w. age×month
15-19 20-24 15-19 20-24 15-19 20-24 15-19 20-24
(1) (2) (3) (4) (5) (6) (7) (8)
Year 2020 -0.66* -4.72*** -0.59*** -3.76*** -0.41** -3.42*** -0.51*** -3.48***
(0.36) (0.91) (0.21) (0.60) (0.17) (0.56) (0.17) (0.50)
Year 2021 0.09 -1.21* 0.52*** -1.63*** 0.70** -1.56*** 0.71*** -1.43***
(0.34) (0.64) (0.18) (0.33) (0.27) (0.34) (0.21) (0.35)
Year 2022 0.04 -1.97** 0.70*** -1.83*** 1.08*** -1.81*** 1.35*** -1.18**
(0.56) (0.94) (0.18) (0.52) (0.37) (0.53) (0.31) (0.57)
UI Expiration 0.20 0.52 -0.21 1.18** -0.06 1.49*** -0.36 0.99**
(0.41) (0.66) (0.16) (0.53) (0.30) (0.39) (0.28) (0.48)
Covid case 0.09 0.53
(0.11) (0.39)
Covid death -22.08* -13.41
(12.30) (27.07)
Y2020=Y2021 0.02 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Y2020=Y2022 0.13 0.02 0.00 0.02 0.00 0.03 0.00 0.00
Y2021=Y2022 0.89 0.21 0.15 0.61 0.13 0.53 0.00 0.54
Obs 60906 50620 560397 360175 454649 416060 454649 416060

Note: *** p<0.01, ** p<0.05, * p<0.1. Table presents estimates of 𝛽𝛽s from equation (3). Year 2020 is a 0-
1 indicator for Mar 2020-Dec 2020. Year 2021 is a 0-1 indicator for the entire year of 2021. Year 2022 is
a 0-1 indicator for Jan 2022-Oct 2022. The last three rows report the T-test for equality of the estimated
coefficients for two years. COVID case (death) records monthly rate of COVID cases (death) per 1000
population in the state and are from the Johns Hopkins Coronavirus Resource Center. The pre-pandemic
average usual working hour is 6.50 for 15-19 MIS 2+ sample (column 1) and 22.81 for 20-24 MIS 2+
sample (column 2).

25
Usual work hours
2

-2

-4

-6

-8
19Jan 20Jan 21Jan 22Jan

15-19 20-24

Figure 1:

Effect of the pandemic on usual work hours

Note: Figure shows estimated coefficients and 95% CIs on month indicators in Eq. 1.

26
Employed
.1

-.1

-.2

-.3
19Jan 20Jan 21Jan 22Jan

15-19 20-24

Figure 2:

Effect of the pandemic on employment

Note: Figure shows estimated coefficients and 95% CIs on month indicators in Eq. 1.

27
Appendix Table 1: Summary Statistics

Age group: 15-19 20-24
(1) (2) (3) (4) (5) (6)
All Before After All Before After
Age 16.92 16.92 16.92 21.68 21.70 21.65
Female 0.49 0.49 0.49 0.45 0.45 0.46
Black 0.15 0.15 0.15 0.16 0.16 0.16
Other race 0.11 0.11 0.12 0.10 0.10 0.10
Hispanic 0.23 0.22 0.24 0.21 0.21 0.22
In central city 0.25 0.25 0.25 0.30 0.30 0.29
Outside central city 0.48 0.48 0.48 0.44 0.44 0.44
Hours, all occupations 6.11 5.94 6.38 21.09 21.49 20.45
-Food preparation and
serving 1.54 1.48 1.63 2.76 3.00 2.37
-Sales and related 1.20 1.23 1.17 2.99 3.13 2.78
-Other occupations 3.37 3.24 3.57 15.34 15.37 15.29
Employment rate 0.25 0.25 0.25 0.61 0.63 0.58
Work full-time 0.06 0.06 0.07 0.36 0.36 0.35
Work part-time 0.19 0.19 0.19 0.25 0.26 0.24
School attendance 0.60 0.61 0.59 0.41 0.42 0.39
Obs 560397 368450 191947 360175 237084 123091

Note: Data come from the CPS, 2016 Jan-2022 Oct, with 2016 Jan-2020 Feb being the before-pandemic
time period, and 2020 Mar-2022 Oct being the after-pandemic time period.

28
Appendix Table 2: Full-time and Part-time status by school attendance

Month April of 2016-2019 July of 2016-2019
Age group 15-19 20-24 15-19 20-24
(1) (2) (3) (4)
April in school 67% 46% 68% 46%
Full-time status 5% 34% 9% 41%
Part-time status 18% 28% 21% 25%
Full-time|April in school 2% 10% 9% 26%
Part-time|April in school 21% 37% 25% 33%
Full-time|April out school 11% 54% 10% 53%
Part-time|April out school 12% 21% 11% 18%
Obs 29687 19606 6147 3810

Note: The statistics for July are based on a subsample who are observed in April of the same year.

29
Appendix Table 3: Effects of the pandemic on other age groups

Usual Work Hours
Childless, no college degree all
25-54 55-64 65-74 25-54 55-64 65-74
(1) (2) (3) (4) (5) (6)
Year 2020 -3.52*** -1.58*** -0.83*** -2.64*** -1.77*** -0.87***
(0.44) (0.29) (0.16) (0.31) (0.23) (0.14)
Year 2021 -2.53*** -1.39*** -0.61*** -1.77*** -1.06*** -0.78***
(0.28) (0.24) (0.16) (0.14) (0.13) (0.14)
Year 2022 -2.16*** -0.84** -0.49 -1.19*** -0.59** -0.63**
(0.31) (0.41) (0.30) (0.17) (0.23) (0.27)
UI Expiration 1.24*** 0.31 0.12 0.70*** 0.36* 0.27
(0.25) (0.27) (0.22) (0.13) (0.18) (0.20)
Y2020=Y2021 0.03 0.55 0.22 0.01 0.01 0.56
Y2020=Y2022 0.01 0.15 0.33 0.00 0.00 0.46
Y2021=Y2022 0.05 0.08 0.60 0.00 0.01 0.49
Obs 909854 629445 565765 3253047 1255804 1019358
Pre-pandemic mean 29.26 21.96 7.21 32.75 25.35 9.03
Note: *** p<0.01, ** p<0.05, * p<0.1. Table presents estimates of 𝛽𝛽s from equation (3). Year 2020 is a 0-
1 indicator for Mar 2020-Dec 2020. Year 2021 is a 0-1 indicator for the entire year of 2021. Year 2022 is
a 0-1 indicator for Jan 2022-Oct 2022. The last three rows report the T-test for equality of the estimated
coefficients for two years.

30
Hours in food preparation and serving occupations
1

-1

-2
19Jan 20Jan 21Jan 22Jan

15-19 20-24

Appendix Figure 1:

Effect of the pandemic on usual work hours in food preparation and serving occupations

Note: Figure shows estimated coefficients and 95% CIs on month indicators in Eq. 1.

31
Hours in sales and related occupations
1

-.5

-1

19Jan 20Jan 21Jan 22Jan

15-19 20-24

Appendix Figure 2:

Effect of the pandemic on usual work hours in sales and related occupations

Note: Figure shows estimated coefficients and 95% CIs on month indicators in Eq. 1.

32
Hours in other occupations
2

-2

-4

-6
19Jan 20Jan 21Jan 22Jan

15-19 20-24

Appendix Figure 3

Effect of the pandemic on usual work hours in other occupations (not food preparation and
serving or sales and related)

Note: Figure shows estimated coefficients and 95% CIs on month indicators in Eq. 1.

33
Work Full-time
.05

-.05

-.1

-.15

-.2
19Jan 20Jan 21Jan 22Jan

15-19 20-24

Appendix Figure 4

Effect of the pandemic on full-time employment

Note: Figure shows estimated coefficients and 95% CIs on month indicators in Eq. 1.

34
Work Part-time
.05

-.05

-.1

-.15
19Jan 20Jan 21Jan 22Jan

15-19 20-24

Appendix Figure 5

Effect of the pandemic on part-time employment

Note: Figure shows estimated coefficients and 95% CIs on month indicators in Eq. 1.

File and source

File: recovery-from-the-covid-19-recession-uneven-effects-among-young-workers.pdf
Size: 403,417 bytes
SHA-256: 764fb95af6db1ac53b5b5fbf7fe21795a9fe175ed13d239e022615c41db8b445
Our copy: recovery-from-the-covid-19-recession-uneven-effects-among-young-workers.pdf
Original: openalex.org

Recovery from the COVID-19 Recession Among Young Workers

Summary

Full text

File and source

Related records