Which Workers Bear the Burden of Social Distancing? (Mongey, Pilossoph and Weinberg, NBER Working Paper 27085)
Summary
NBER Working Paper No. 27085, Which Workers Bear the Burden of Social Distancing?, by Simon Mongey, Laura Pilossoph and Alex Weinberg, dated May 2020, Revised March 2021. The abstract states that the authors use O*NET data to construct two measures of an occupation's potential exposure to social distancing measures, the ability to conduct the job from home and the degree of physical proximity to others the job requires. It states that workers in low-work-from-home and high-physical-proximity jobs are more economically vulnerable, being less educated, of lower income, with fewer liquid assets relative to income and more likely renters. It also states that these workers experienced greater declines in employment during the pandemic and compares the aggregate employment losses with those of the 2008 recession.
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
- Issuer
- Source documents
- Document type
- Which Workers Bear The Burden Of Social Distancing Policies
- Case
- Which Workers Bear The Burden Of Social Distancing Policies
Cite this document
“Which Workers Bear the Burden of Social Distancing? (Mongey, Pilossoph and Weinberg, NBER Working Paper 27085).”
Original at www.nber.org. Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/which-workers-bear-the-burden-of-social-distancing-policies/
, SHA-256 10824914fa3bb4b2d2bde3a3560aa029e7d9a238c3285e9b366fec84a688495f.Full text
NBER WORKING PAPER SERIES
WHICH WORKERS BEAR THE BURDEN OF SOCIAL DISTANCING?
Simon Mongey
Laura Pilossoph
Alex Weinberg
Working Paper 27085
http://www.nber.org/papers/w27085
NATIONAL BUREAU OF ECONOMIC RESEARCH
1050 Massachusetts Avenue
Cambridge, MA 02138
May 2020, Revised March 2021
Thanks to Gianluca Violante and Greg Kaplan for making available their codes. Our measures at
the three digit occupation level are available on our websites. The views expressed in this study
are those of the author and do not necessarily reflect the position of the Federal Reserve Bank of
New York, the Federal Reserve System, or the National Bureau of Economic Research.
Replication code is available on the authors' websites.
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 They have not been
peer-reviewed or been subject to the review by the NBER Board of Directors that accompanies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.
© 2020 by Simon Mongey, Laura Pilossoph, and Alex Weinberg. 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.
Which Workers Bear the Burden of Social Distancing Policies?
Simon Mongey, Laura Pilossoph, and Alex Weinberg
NBER Working Paper No. 27085
May 2020, Revised March 2021
JEL No. E24,J01,J22
ABSTRACT
Using data from O*NET, we construct two measures of an occupation’s potential exposure to
social distancing measures: (i) the ability to conduct that job from home and (ii) the degree of
physical proximity to others the job requires. After validating these measures with comparable
measures from ATUS as well as realized work-from-home rates during the pandemic, we employ
the measures to study the characteristics of workers in these types of jobs. Our results show that
workers in low-work-from-home and high-physical-proximity jobs are more economically
vulnerable across various measures constructed from the CPS and PSID: they are less educated,
of lower income, have fewer liquid assets relative to income, and are more likely renters.
Consistent with the idea that high physical proximity or low work-from-home occupations were
more exposed to the Coronavirus shock, we show that the types of workers predicted to be
employed in them experienced greater declines in employment during the pandemic. We
conclude by comparing the aggregate employment losses in these occupations to their
employment losses in the 2008 recession, and evidence that these occupations were
disproportionately exposed to the pandemic shock, and not just comprised of more cyclically
sensitive workers.
Simon Mongey Alex Weinberg
Kenneth C. Griffin Department of Economics 1126 E 59th St
University of Chicago Chicago, IL 60637
1126 E. 59th Street weinberga@uchicago.edu
Chicago, IL 60637
and NBER
mongey@uchicago.edu
Laura Pilossoph
Macroeconomics and Monetary Studies Function
Federal Reserve Bank of New York
33 Liberty Street
New York, NY 10045
pilossoph@gmail.com
1 Introduction
A key response to the Coronavirus pandemic was ‘social distancing’, the reduction of in-person
contact with others. This was reflected in both policy responses - through the shutdown of various
businesses - and in behavioral responses - through the voluntary curtailing of face-to-face activities
(Alexander and Karger (2020),Goolsbee and Syverson (2021)). Such social distancing reduces the
spread of the virus, but can reduce labor demand in occupations that cannot be performed remotely
or require a high degree of physical-proximity.
Understanding which occupations can be performed remotely or require high degrees of physical
proximity is crucial for understanding the economic consequences of the epidemic.1 In particular, to
the extent that workers vary systematically across these jobs, social distancing will have systemati-
cally different effects across individuals. Therefore, understanding how individuals vary across these
occupations is important for policy makers interested in formulating targeted worker assistance pro-
grams. Our paper documents that workers employed in these pandemic-exposed occupations are
disproportionately likely to be economically vulnerable. For example, workers who cannot work
remotely are 40 percentage points more likely to lack a college degree and 23 percentage points
more likely to earn less than the median wage.
Focusing on the US2 , we combine multiple data sources to study how individuals vary across
occupations that differ in the labor demand exposure to social distancing.3 We measure an occupa-
tion’s exposure by (i) a job’s ability to be performed at home and (ii) its required degree of physical
proximity to others. To this end, we merge individual-level data from the Bureau of Labor Statistics’
Current Population Survey (CPS) and the Panel Study of Income Dynamics (PSID) with a version
of the Dingel and Neiman (2020) classification of an occupations’ capacity to work from home as
well as a measure of physical proximity in the workplace. We construct these two measures using
occupation-level data from the Department of Labor’s Occupational Information Network (O∗ NET)
data.4 We show that despite being negatively correlated, some outlier occupations such as those
related to education are both high work-from-home and high physical-proximity, hence relatively
more affected if social distancing policies become targeted.
We validate the measures of work-from-home and physical-proximity using data from the Amer-
1
As an example of the former in practice, our measures are used by Baqaee et al. (2020b) to construct a “GDP-
to-Risk” index which measures which sectors provide the largest improvement to output for a marginal increase in
R0. Indeed, Governor Andrew Cuomo’s policy for NY State consists of a Phase I reopening with Construction and
Manufacturing jobs, which the state views as low risk and highly essential.
2
While our analysis focuses on classifying occupations within the United States, more recent work has broadened
the Dingel and Neiman (2020) analysis to developing countries (Gottlieb et al. (2021)), highlighting that tasks may
be better suited to measure a jobs ability to be done from home rather than an occupation.
3
We do not consider the labor supply effects of policies such as the CARES Act.
4
In these occupation-level data, occupational classifications are finer than those available in the individual-level
data. To make the data conformable we develop a cross-walk that allows us to use the Bureau of Labor Statistics’
Occupation Employment Statistics (OES) to employment weight O∗ NET measures within the coarser occupations
defined in the CPS. Code is available on request.
1
ican Time Use Survey (ATUS) and the CPS.5 The O∗ NET-based work-from-home measure is
designed to capture whether a job could feasibly be done from home and is based on the types
of activities conducted at work (e.g. heavy lifting, working outdoors etc). Nonetheless, we show
that, across occupations, the measure is highly correlated with the share of time working that is
spent at home in ATUS in 2018. Moreover, we show that the O∗ NET physical-proximity measure is
correlated with the reported fraction of time spent working alone in ATUS in 2018. The correlation
of our measure with pre-pandemic measures of actual time spent working from home and the actual
number of people one works with is reassuring. Additionally, using data that has become available
since the start of the pandemic in the supplemental questions to the CPS, we show that workers
employed in jobs that our measures predict can be performed from home were indeed more likely
to telecommute during the pandemic.
With validated occupation-level measures in hand, we present our main results in two steps.
First, we study how individual characteristics of workers vary across these types of occupations.
Our main result is that workers in occupations that are more likely to be affected by social distancing
policies are workers we would consider more economically vulnerable. Workers in these occupations
are less likely to have a college degree and are less likely to have health insurance provided by
their employer. They are less likely to be white, less likely to work at a large firm, and less
likely to be born in the United States. Workers in low work-from-home occupations also have
disproportionately low levels of liquid assets, which is especially important for policies that provide
liquidity to households. We also show that these effects are monotonic: occupations that score
relatively lower (higher) in terms of the work-from-home (personal-proximity) measure, are even
more economically vulnerable.6
Second, we turn to employment outcomes and study employment changes across the February,
April, and August 2020 CPS surveys. Occupations that rank low in the work-from-home mea-
sure and high in the physical-proximity measure experienced larger employment declines relative
to pre-pandemic February-April changes. A direct corollary of our earlier analysis is that more
vulnerable workers did indeed experience larger declines in employment. For example, non-college
educated workers experienced a 15 ppt larger decline in employment relative to those with a college
degree from February to April; this difference decreased to just 10 ppt by August as establishments
reopened. We show that employment losses for workers in low work-from-home and high physical
proximity occupations during the pandemic far exceeded the losses during the 2008 recession. This
suggests that exposure to the COVID shock is partly responsible for the employment outcomes we
document, and not just the fact that more exposed occupations tend to employ more workers that
5
? use Swedish vacancy data to show that job seekers redirect their search towards high work-from-home occu-
pations during the pandemic. Their measure of an occupations’ ability to be done from home is based on the ATUS
data.
6
When we compare the top quartile of occupations by the work-from-home measure, to the bottom quartile of
occupations by the work-from-home measure, we find that the estimated treatment effects are larger. When we
compare the third quartile of occupations by this measure, to the second quartile of occupations by this measure, we
find that the estimated treatment effects are smaller but still statistically significant in all cases.
2
are sensitive to downturns more generally.
Our results have clear implications for economic inequality and public policy responses to the
pandemic. First, our results provide guidance as to how income replacement and liquidity injection
policies may be targeted. Indeed the various programs enacted through the CARES Act may have
stemmed much of the wage losses associated with job loss in these occupations. Second, since
low work-from-home and high physical-proximity workers tend to have lower incomes and lower
liquidity, the marginal social cost of income support is low, while the marginal private benefits
are high. Social benefits are also high: such workers have higher propensities to consume out of
transfers, and high disease transmission risk if they do work. Third, the correlation between low
work-from-home and high physical-proximity jobs creates a double-edged sword for workers. It
induces a correlation between economic risks under tight social distancing and health risks under
relaxed social distancing. Already more economically vulnerable workers are disproportionately
exposed to unemployment now, and infection in the future, suggesting the need for on-going policy
interventions.
Literature. Since the start of the pandemic, the literature surrounding the theory and economic
consequences of social distancing has boomed. On the empirical side, Dingel and Neiman (2020)
use the Occupational Employment Statistics (OES) to ask the important question of what fraction
of employment and income is accounted for by jobs that can be done from home. Leibovici et al.
(2020) conduct a similar analysis, instead considering low physical-proximity occupations rather
than high work-from-home occupations.7 Both use the O∗ NET to classify occupations, and then
employment and income data from the OES to study the geographic distribution of employment
and income accounted for by types of jobs. Our focus here is on understanding the characteristics
of the underlying workers that comprise employment in these jobs, validating the measures by
showing they are consistent with measures from other datasets, and verifying that they are indeed
correlated with post-outbreak outcomes.8 This requires integrating the O∗ NET data with data
containing worker characteristics, such as the CPS and the PSID.
While we validate our measures with both pre-pandemic time use data and data on telecom-
muting collected during the pandemic, several authors have conducted their own surveys to collect
information on teleworking. For example, Adams-Prassl et al. (2020) run a survey in the UK, US
and Germany, and find that jobs with mostly WFH tasks saw smaller declines in wages and em-
ployment. Similarly, Bick et al. (2020) conduct a survey in the US and confirm that 35 percent of
workers telecommuted in May 2020.
Finally, on the theory side, several papers have studied the macroeconomic and distributional
7
St. Louis Federal Reserve, On the Economy blog:
https://www.stlouisfed.org/on-the-economy/2020/march/social-distancing-contact-intensive-occupations
8
Lekfuangfu et al. (2020) also use O∗ NET data to characterize jobs, but appeal to factor analysis to define high
and low work from home or physical proximity jobs. Applying their measures to Thailand, they also study worker
characteristics within different types of jobs, with a focus on how couples are sorted into these jobs.
3
consequences of the pandemic using our data on occupational characterizations as inputs. For exam-
ple, Akbarpour et al. (2020) use a heterogeneous agent model to evaluate targeted social distancing
policies, and discipline their model using our data on an industry’s WFH ability. Similarly, Ka-
plan et al. (2020) develop a macro-SIR (“susceptible, infected, recovered”) model with occupational
heterogeneity disciplined by our estimates. Baqaee et al. (2020a) also develop an epidemiological
model disciplined by our data to think about various reopening strategies.
Overview. Section 2 describes how we construct our measures of work-from-home and physical-
proximity using the O∗ NET and OES data. We compare the two measures across occupations,
and validate each against comparable measures constructed from the pre-pandemic ATUS data.
We further validate the measures using realized work-from-home rates from the CPS supplement
questions introduced during the pandemic. Section 3 integrates the CPS and PSID data and gives
our main results, which are summarized in Figure 4. Section 4 shows how individuals employed in
occupations characterized by their work-from-home and physical-proximity measures fared in over
the implementation of social-distancing. Section 5 concludes.
2 Low work-from-home and high physical-proximity jobs
We now describe the construction of our work-from-home and personal-proximity measures, discuss
how the measures compare across occupations, and validate both measures against ATUS data and
realized work-from-home behavior during the pandemic. A key data contribution is our simple,
portable procedure for aggregating O∗ NET data and Dingel and Neiman (2020)’s telework measure
from the SOC occupational classification to the OCC occupation classification system used by the
Census Bureau. More details regarding the data construction are relegated to Appendix A.
2.1 Construction of pandemic exposure measures
We use O∗ NET data on work activities to construct two measures of an occupation’s exposure to
social distancing. We sign these measures in terms of the expected negative economic impacts of
the pandemic: (i) low work-from-home (LW F H j ), and (ii) high physical-proximity (HP P j ).
Our measure HP P j is simply the O∗ NET variable which measures the physical-proximity re-
quired by an occupation on a scale of 1 − 5. Occupations which require workers to be in very
close physical proximity to others receive scores of 5.9 We define the binary variable HP P ∗j to
take a value of 1 for occupation j if HP P j is above the employment-weighted median across OCC
occupations of physical proximity (= 3.6) and a value of zero if it is below. For certain figures, we
aggregate HP P j to the 2 digit OCC level and rescale so that HP P j ∈ [0, 1].
Our measure of work-from-home, a modification of the telework measure developed by Dingel
and Neiman (2020), captures an occupation’s ability to be performed remotely. We ask whether
9
The exact question and the possible answers can be found here: https://www.onetonline.org/find/descriptor/result/4.C.2.a.3.
4
(i) Professional, management, technology; (ii) Services (iii) Clerical, sales; (iv) Production, operators
1.0 Professional, management, tech. 1.0 Clerical, sales Install/Maintenance/Repair
Construction/Extraction
Services Production, operators
0.8 0.8
Low work−from−home >> Low work−from−home >>
Healthcare supp.
Production Material moving
Building maintenance Protection services
Transport
0.6 0.6
Food prep.
Healthcare tech. Farm/Fish/Forest
0.4 Personal care 0.4
0.2 0.2 Sales
Science Community/Social
Entertainment/Media
Management Architecture/Engineering Office/Admin
Business/Financial
Computer/Math
0.0 Legal Education 0.0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
High physical−proximity >> High physical−proximity >>
Figure 1: Occupations by Work-from-home and Physical-proximity (2 digit, Census OCC)
Notes: This figure compares groups of 2 digit OCC code occupations. We split the data into panels A and B only for
readability, so that occupation titles can be included. To construct this figure, we employment-weight using the OES
to aggregate LW F H j and HP P j to the 2 digit level. The gray line plot fitted values and 95% confidence intervals
from an employment-weighted linear regression across 2 digit occupations. Occupations above the red-dashed line have
LW F Hj∗ = 1, and account for half of employment. Occupations to the right of the blue-dashed line have HP Pj∗ = 1,
and account for half of employment.
each occupation is intensive in 17 O∗ NET work activities.10 Our continuous measure LW F H j is
a tally ∈ [0, 17] of the number of in-person activities required by the occupation. An occupation
which requires workers to perform many in-person activities (i.e. LW F H j is large) is less able to be
done at home. Our binary variable LW F H ∗j takes a value of 1 if LW F H j is above the employment
weighted median (= 2) and a value of 0 otherwise. For certain figures, we aggregate LW F H j to
the 2 digit OCC level and rescale so that LW F H j ∈ [0, 1].
In our discussion, we will use occasionally use HP P and ‘high PP’ to refer to HP P ∗j = 1
occupations and LP P and ‘low PP’ to refer to HP P ∗j = 0 occupations. Similarly, LW F H and ‘low
WFH’ refer to LW F H ∗j = 1 occupations while HW F H and ‘high WFH’ represent LW F H ∗j = 0
occupations.
2.2 Which jobs are low work-from-home and high physical-proximity?
Figure 1 shows how occupations - aggregated to the 2-digit OCC level - vary across these two metrics,
and where our cut-offs lie for the binary measures.11 Unsurprisingly, there is a strong positive
correlation between low work-from-home and high physical-proximity occupations.12 Typical office
10
Examples include: (i) Working outdoors and (ii) Repairing and maintaining mechanical equipment. The full list
can be found in Appendix A.
11
For readability of this figure, we employment-weight using the OES to aggregate LW F H j and HP P j to the 2 digit
M in M ax M in
level. We then linearly transform each measure X j using its minimum and maximum: (X j − X j )/(X j − X j ).
12
In Appendix Table A4, we provide a full list of 2-digit Occupations, their rankings across these measures, and
their employment shares.
5
jobs in financial services or the legal profession have few of the features that make it unamenable
to being done from home. There is also little work done within arm’s length in these jobs. On the
other hand, construction, material moving, and healthcare jobs are low work-from-home and high
physical proximity.
A number of occupations stand out as deviations from this pattern. Education jobs require
close physical-proximity, but little of the features that would prevent the job being conducted
at home. Under broad social-distancing, workers in these jobs can successfully stay employed
while operating from home, which has indeed occurred through virtual teaching. Agricultural jobs
(Farm/Fish/Forest), meanwhile, may pose lower contagion risk due to low physical-proximity, but
are difficult to be done from home. Such jobs may be punished somewhat unduly by indiscriminate
social-distancing.
Note that while Figure 1 shows occupations at the aggregated 2 digit OCC level, many more
narrow 3 digit OCC occupations differ from their group. Consider for example the 2 digit OCC
Entertainment/Media. The broad occupation is HP P ∗ = 0 even while a 3 digit OCC Dancer is
LP P ∗ = 1. Figure B1 plots all 3 digit occupations.
Figure 1 and Table A4 rank 2 digit OCC occupations along LW F H j and HP P j . The continuous
measures can be downloaded at both the 3 digit and 2 digit OCC levels online.13
2.3 Validation I - Comparison to ATUS
We validate our occupation-level pandemic-exposure measures using the behavior of workers in
those occupations in the 2018 American Time Use Survey (ATUS). The ATUS reports where and
with whom individuals do various activities. To validate our measure of whether an occupation can
be performed from home, we compare the measure to the share of an occupation’s work hours that
are spent at home. We validate our physical-proximity measure against an imperfect proxy, the
share of an occupation’s work hours that are spent alone.14
Both O∗ NET measures are negatively correlated with their ATUS counterpart, validating the
measures. At the 2 digit level, the ATUS share of work hours at home and LW F H j have a
correlation of −0.80. The physical proximity measure is less tightly linked, with a correlation of
−0.56. The looser fit is to be expected given that the ATUS measures whether one is working
with co-workers while the O∗ NET measure uses information on how physically close workers are to
others, including customers.
Figure 2 depicts these correlations graphically. It also provides preliminary evidence on the
distributional effects of social distancing. Workers in professional services jobs (blue markers)
13
Download our measures here: https://github.com/simonmongey/Mongey_Pilossoph_Weinberg/tree/master/
files
14
We use the question in the ATUS “who” file which asks - for each activity the respondent recorded - “Who was
in the room with you/Who accompanied you?” for the measure of hours working spent alone. We use the question
from the interview file which asks “where were you during this activity?” for the measure of hours spend working at
home.
6
A. Share of working time spent working at home B. Share of working time spent working alone
0.5 Professional, manage., tech.
Share of work hours at home (ATUS) Share of work hours alone (ATUS)
Services
0.5
0.4 Clerical, sales
Production, operators
0.4
0.3
0.3
0.2
0.1 0.2
0.0 0.1
0.0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
Low work−from−home >> (O*NET) High physical−proximity >> (O*NET)
Figure 2: Comparing work-from-home and physical-proximity measures to ATUS
Notes: Panel A compares the fraction of individuals reporting that they can work from home in the ATUS against
the O∗ NET WFH measure. Panel B compares the physical proximity measures constructed from the two datasets.
The share of adjusted work hours accounts for the fact that respondents can answer that they are with multiple
individuals while performing a particular activity. Fitted values are from employment-weighted linear regressions,
and display 95 percent confidence intervals for the conditional expectation of the dependent variable.
already spend a significant fraction of time working from home and more time working alone.
These types of workers—usually higher income and college educated—will be less likely impacted
by social distancing. We study this in detail using individual-level data in Section 3.
2.4 Validation II - Measures from the CPS Covid-19 Supplement
As a second validation that the pandemic exposure measures are useful, we compare them with
realized work-from-home behavior during the pandemic. Starting in May 2020, a series of supple-
mental questions were added to the CPS to understand the labor market and health impacts of the
pandemic. In this Covid Module, the CPS asked respondents “At any time in the last 4 weeks, did
you telework or work at home for pay because of the coronavirus pandemic?” As shown in Table B5,
during the five month period for which data is currently available (May - November, 2020) individ-
uals employed in HW F H ∗ occupations were 35 percentage points more likely to report teleworking
than those employed in LW F H ∗ occupations. During the same period, individuals employed in
LP P ∗ occupations were 18 percentage points more likely to work-from-home than those employed
in HP P ∗ occupations.
Figure 3 validates our measures against this novel data graphically. Panel A (Panel B) of Figure
3 shows the relationship between our LW F H (HP P ) measures and the share of respondents that
reported teleworking in the May survey. Our work-from-home measure is the stronger predictor of
whether or not workers teleworked during the Covid-19 pandemic. Our regression in Table B5 of
an indicator of teleworking on LW F H ∗ yields an R2 = 0.16 while a similar regression on HP P ∗
yields R2 = 0.04. Table B4 shows the percent of workers in each occupation type that reported
7
A. Work−from−home B. Physical−proximity
Non−essential
Share teleworking due to Covid (May) Share teleworking due to Covid (May)
Essential
0.80 0.80
0.60 0.60
0.40 0.40
0.20 0.20
0.00 0.00
−0.20 −0.20
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
Low work−from−home >> High physical−proximity >>
Figure 3: Comparing work-from-home and physical-proximity measures to CPS Covid Module
Notes Data is aggregated to the 2 digit occupation level. Circle sizes reflect employment in each occupation. Panel
A compares the fraction of individuals in May 2020 reporting that they teleworked due to the Covid-19 pandemic
against the O∗ NET WFH measure. Panel B compares the physical-proximity measure with the share of workers who
report that they teleworked. Fitted values are from employment-weighted linear regressions of occupations marked
as non-essential.
working-from-home each month.
3 Characteristics of workers in exposed jobs
We now compare the characteristics of workers employed in low work-from-home LW F Hj∗ = 1 and
high physical-proximity HP Pj∗ = 1 occupations with workers in LW F Hj∗ = 0 and HP Pj∗ = 0
occupations respectively.
We merge our validated measures with worker-level data in the March CPS and PSID. We
construct our PSID sample following Kaplan et al. (2014) and our CPS sample following Heathcote
et al. (2010).15 The key finding is that workers employed in social-distancing exposed occupations
are disproportionately low income, low education, and economically vulnerable more generally.
3.1 Approach
Our approach is simple and designed to be easily interpretable. Let yij be a binary characteristic of a
worker i employed in occupation j last year. For simplicity, we work work with binary variables. As
an example, we construct the binary variable ‘below median income’ from the continuous variable
‘wage’. We regress each worker characteristic yij on both pandemic-exposure measures. Using
15
We follow Heathcote et al. (2010)’s sample selection criteria for their Sample C which is as follows. We construct
wages by dividing total wage and salary income by annual hours worked. Annual hours are the product of weeks
worked last year and usual weekly hours. We restrict our sample to individuals aged 25-65 who work at least 260
hours (equivalent to working a month of 8 hour days)paid wages at least half of the Federal minimum wage.
8
LWFH,; as an example, we estimate:
Yij = Ay + ByLW FH; + ej. (1)
This gives the sample moment:
By = E [yij|LWFH® = 1] —E [yij| LW FH = 0)
where E is the sample mean. Since y;; is binary, By is simply the fraction of workers for which
Yij = 1 in low work-from-home occupations, relative to the fraction of workers for which y;; = 1 in
high work-from-home occupations. Our estimate By is a measure of how disproportionately y;; = 1
workers in low work-from-home occupations are.
We estimate (1) for the individual characteristics listed below. In each case we assign y;; = 1 to
the individuals with the characteristic most related to being in a low work-from-home occupation.
- Demographics. (i) Non-white, (ii) No college degree, (iii) Age below 50, (iv) Male, (v) Single, (vi)
Born outside USA, (vii) Non-US citizen, (viii) Rent their home
- Work. (i) No healthcare provided by employer,!® (ii) Employed at a small firm (< 500 employees),
(iii) Part-time employed
- Income. (i) Below median wage (ii) Experienced a spell of unemployment in the last year. (iii)
Hand-to-mouth (iv) Poor hand-to-mouth!”
3.2 Results
Our main results consist of plotting By in Figure 4.!° Clearly By € [—1,1] and takes the maximum
value of 1 when y;; = 1 for all individuals for which LWF'H; = 1, and y;; = 0 for all individuals
for which LW FH; = 0. Comparing estimates across measures y and y’, a higher value of By > By
can be interpreted as
“Workers in occupations for which LW F H; = 1 are relatively more different from workers in
occupations for which LWFH; = 0 along dimension y than along dimension y!”.
In Figure 4A, we plot the estimates for each of these characteristics for the low work-from-home
regression, ordering these attributes from the highest to the lowest point estimate. Figure 4B repeats
the exercise for the high personal-proximity regression. For most of the individual characteristics,
the results for high work-from-home occupations and low physical-proximity occupations are the
same in terms of their sign, as is evident from the fact that most of the dots are to the right of
16We set the indicator for employer provided healthcare to 1 if the employer pays for any part of the individual’s
health insurance premiums.
‘Households are hand-to-mouth when they are liquidity-constrained. See Kaplan et al. (2014) for details on
construction using the PSID.
Table Bl and Table B2 further decompose these results; providing the moments E [yi |LWE Hj = 0] and
E |yi;|LWFH = 1], and similarly for HPP.
A. Low work−from−home jobs B. High physical−proximity jobs
No college degree Below median wage
Below median wage No college degree
Male No employer healthcare
No employer healthcare Part−time employed
Rents home Single
Born outside US Rents home
Single Age <50
Small firm (<500 emp.) Non−white
Non−US citizen Born outside US
Non−white Unemployed at all last year
Part−time employed Non−US citizen
Unemployed at all last year Small firm (<500 emp.)
Age <50 Male
Hand to mouth (PSID) Hand to mouth (PSID)
Poor hand to mouth (PSID) Poor hand to mouth (PSID)
0.00 0.10 0.20 0.30 0.40 −0.10 0.00 0.10 0.20
More likely >> << Less likely , More likely >>
Figure 4: Characteristics of workers in Low Work-from-home and High physical-proximity jobs
Notes This figure plots estimates of βby for 10 characteristics y from regressions in which LW F H ∗j ∈ {0, 1} is the
independent variable (Panel A), and in which HP P j ∈ {0, 1} is the independent variable (Panel B). If x% of workers
in high work-from-home occupations have, for example, no college degree, then Panel A shows that (x+38)% of workers
in low work-from-home occupations have no college degree. A high value of βby therefore means that workers in low
work-from-home occupations are more likely than workers in high work-from-home occupations to be in the category
listed on the vertical axis. Point estimates are given by the circle markers, and 95 percent confidence intervals are
given by the lines through each marker. All blue results are derived from the CPS, red results are derived from the
PSID.
zero. For example, workers in both high physical-proximity occupations and low work-from-home
occupations are less likely to have a college degree than workers in low physical-proximity and high
work-from-home occupations, respectively. The results are less stark for the high physical proximity
occupations, as the magnitudes of the coefficients are usually smaller.
Occupations which cannot be performed from home or that have high physical proximity re-
quirements feature workers that, by all measures, are economically more vulnerable. Workers in
low WFH (high PP) occupations are 40 (12) percentage points less likely to have a college degree
and 23 (19) ppt more likely to be below median income. They are more likely to rent rather than
own their homes and so are less likely to be in positions to take advantage of interest rate cuts, and
have fewer collateralizable assets to borrow against to compensate for earnings losses. Additionally,
those in low WFH occupations are more likely to work in smaller firms (though this is not the case
for workers in high physical proximity occupations), which are on average less financially robust
and so more likely to suffer from the financial effects of the crisis (Chodorow-Reich, 2014).
Workers in low WFH and high PP jobs are also less likely to have access to informal insurance
channels that may help them weather the crisis. They are less likely to be married, which can
diversify household income against individual income risk. They are less likely to be US citizens or
born in the US, which may lead to less family support, as well as restricting access to emergency
10
government programs. Finally, they are more likely to have unstable employment; they are less
likely to be employed full-time and more likely to have recently experienced unemployment.
Availability of healthcare is obviously a key insurance mechanism in a pandemic. Workers in
low work from home occupations and high physical proximity occupations are less likely to have any
employer-subsidized healthcare. However, we find that the age of workers across these high- and
low- work-from-home occupations does not systematically differ. Given that the mortality rate for
those with COVID-19 is significantly higher for older individuals,19 this means that workers across
the different types of jobs have the same fundamental health risks as they relate to age, but those
in low WFH or high PP jobs are less likely to have the health insurance to provide for them in the
case of infection.
We expect that low access to liquid savings will compound the economic consequences of job loss
or reduction in hours, and the health consequences of infection. To understand whether workers in
low work-from-home jobs have disproportionately lower levels of liquid savings we add data from
the PSID and construct measures of whether a household is hand-to-mouth following Kaplan and
Violante (2014).20 Hand-to-mouth households are households with liquid assets that are less than
half of one month’s income.
The results are depicted in red in Figure 4. We find that households in which the highest
earner is employed in a low work-from-home or high physical-proximity job are disproportionately
hand-to-mouth. Conditional on being hand-to-mouth, households may be poor-hand-to-mouth
or wealthy-hand-to-mouth depending on whether they have positive or negative net-assets, respec-
tively. Conditional on being hand-to-mouth, workers in jobs most likely affected by social distancing
policy are disproportionately poor-hand-to-mouth.
The magnitudes of the point estimates are economically significant. Hand-to-mouth low work-
from-home households are 10 ppt more likely to be poor hand-to-mouth than hand-to-mouth high
work-from-home households. To put this in perspective we could compare this to how, as households
age, the composition of hand-to-mouth households shifts from poor- to wealthy-. Starting at age
30, one would need to move all the way up to age 50—a period of high income growth—in order
to obtain a 10 percent decline in the fraction of hand-to-mouth households that are poor hand-to-
mouth (Kaplan et al., 2014, Figure 6). Despite not being significantly younger, low work-from-home
households have finances as if they are twenty years further back in their lifecycle.
The results differ across these two occupation exposure measures most sharply for sex. Individ-
uals in occupations that score highly in terms of work-from-home are more likely to be male, while
individuals in occupations that have high physical-proximity are more likely to be female. This
relates to the earlier example of Education jobs from Figure 1, which are female-dominated. Taking
these results at face value, female workers may be relatively less affected by universal social dis-
tancing measures, but could be relatively more affected if restrictions or behavior responds around
19
See https://www.cdc.gov/nchs/nvss/vsrr/covid19/index.htm
20
We use the code made publicly available from Kaplan et al. (2014).
11
occupations with higher personal-proximity.
Finally, a policy maker may not be interested in programs targeted below and above the median
of the indexes we create since they rule in too many individuals. We therefore verify that if we make
more extreme comparisons using the tails of our measures, then we get more extreme results. Figure
B3 in Appendix B compares the lower quartile to the upper quartile (dropping the middle quartile,
in red), and the second quartile to the third quartile (dropping the upper and lower quartile, in
green). When we compare workers in very low work-from-home occupations to workers in very
high work-from-home occupations (in red), the coefficients are uniformly larger in magnitude. For
example, workers in the lowest quartile of work-from-home occupations are nearly 50 ppt more likely
to not have a college degree than workers in the highest quartile. Targeting policies into the lower
tail of the distribution is thus both cheaper (lower incomes to replace) and more effective (lower
resources initially available).
Taken together, the evidence shows that those in low work from home occupations and high
physical proximity occupations are less prepared to weather the economic and hardship induced by
the Covid pandemic. Moreover, the correlation between low work-from-home and high physical-
proximity jobs discussed in Section 2 creates a double-edged sword. It induces a correlation in
economic risks due to policy and health risks due to transmission of Coronavirus. More vulnerable
workers are therefore relatively more exposed to both.
Joint examination of work-from-home and physical proximity. Since LW F H j and HP P j
are correlated, perhaps our results are best explained by one measure and not the other. We repeat
our analysis in Figure B2 and B3 jointly examining HPP and LWFH. For example, the left panel
of Figure B2 depicts the coefficient βby from the same WFH regression (1), but conditioning on the
HPP status of the job. The blue dots therefore represent how much more likely workers are to have
attribute y if they are in low WFH jobs, conditional on working in a high PP job. Similarly, the
red crosses represent how much more likely workers are to have attribute y if they are in low WFH
jobs, conditional on working in a low PP job.
Conditional on working in a high physical proximity occupation, the characteristics of workers in
low and high work from home jobs are similar, with a couple of notable exceptions. First, regardless
of physical proximity, workers in low work from home jobs are disproportionately male. Second,
conditional on being in a HPP occupation, low work from home workers disproportionately do not
have a college degree (panel A). As expected, this implies that if instead we first condition on
low or high work from home, then compare workers in low and high physical proximity jobs, the
composition of workers is relatively similar (panel B).
12
4 Employment during the epidemic
We now use the data available since the start of the epidemic to demonstrate that pandemic-exposed
occupations experienced larger employment losses. Our main results in Figure 4 suggest that workers
associated with low work-from-home jobs should expect to see larger employment losses. Indeed,
these workers experienced larger declines in employment during the early months of the outbreak.
4.1 Employment losses by occupation
Excess employment losses from February to April of 2020 show a clear pattern: occupations with low
work-from-home and high physical-proximity scores had relatively larger declines in employment
compared to occupations with high WFH scores and low PP scores, respectively. Jobs that are more
easily done at home and involve lower physical proximity were more likely to remain intact. We show
that this is the case using CPS occupational employment data.21 To account for seasonal factors and
long run trends in occupation employment changes, we construct excess employment losses by taking
the log change in employment from February to April 2020 and subtracting the average February
to April log change in employment from 2010-2019. Figure 5 compares the relationship between
this excess decline in employment against LW F H j (Panel A) and HP P j (Panel B), showing that
low work-from-home jobs and high physical proximity jobs experienced larger excess employment
losses.
Comparing Figure 5 panels A and B of reveals that physical-proximity is the stronger predictor
of employment losses during the initial period of the pandemic from February-April. This may
be because high physical-proximity jobs such as food service or personal care are associated with
infection exposure for the consumer in addition to the worker. By August, however, the relationship
between physical-proximity and employment losses flattened as social distancing mandates expired
and some individuals chose to return to purchasing in-person services such as restaurant meals.
Employment losses remained larger in low work from home occupations which are less sensitive to
demand-side concerns.
Essential workers. An important exception to this relationship, as expected, are those jobs
deemed essential by public policy. These essential occupations are unlikely to have employment
losses that correlate with the ability to telework, or with whether the job entails high physical prox-
imity with others. For example, front line medical workers have very low work-from-home measures
(healthcare supplemental workers have a LW F H j index of around 0.8) and high physical-proximity
measures (healthcare supplemental workers have an HP P j index of 1.0). Because healthcare is
considered essential, workers in these occupations have not experienced the dramatic employment
losses implied by their physical-proximity and work-from-home scores. We use industry data cre-
21
Summary statistics are reported in Table A1
13
Figure 5: Employment declines by occupations and worker characteristics
A. Work−from−home (Feb−Apr) B. Physical−proximity (Feb−Apr)
Relative change in Feb−Apr log employment Relative change in Feb−Apr log employment
0.00 0.00
−0.20 −0.20
−0.40 −0.40
−0.60 Non−essential −0.60
Essential
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
Low work−from−home >> High physical−proximity >>
C. Work−from−home (Feb−Aug) D. Physical−proximity (Feb−Aug)
Relative change in Feb−Aug log employment Relative change in Feb−Aug log employment
0.00 0.00
−0.20 −0.20
−0.40 −0.40
−0.60 Non−essential −0.60
Essential
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
Low work−from−home >> High physical−proximity >>
Notes These figures plots employment changes by 2-digit OCC occupation against LW F H j (Panel A) and HP P j
(Panel B). Employment change is the Feb-April log change in employment in 2020 of each occupation, relative to the
average Feb-April log change in employment over 2010-2019 for the occupation. Panel C and Panel D are similarly
plotted for Feb-August log changes in employment. Occupations marked with red diamonds are defined as “essential”
using the grouping from Tomer and Kane (2020). Fitted values are from an employment-weighted linear regression
estimated on non-essential occupations, and gives 90 percent confidence intervals for the conditional expectation of
the dependent variable. Data is from the CPS.
ated by Tomer and Kane (2020), that categorizes certain 4-digit NAICS industries as “essential”.22
For each 2 digit occupation we use the 2018 OES to calculate the share of employment in essential
industries, and categorize an occupation as essential if more than 75 percent of employment work in
an essential industry. Occupations that meet this criterion are depicted in red in Figure 5. Among
these occupations there is no significant relationship between the WFH measure and employment
22
Tomer and Kane (2020) use job descriptions from the government statement which announced guidelines for
categorizing essential jobs. The text for this document can be found at https://www.cisa.gov/publication/guidance-
essential-critical-infrastructure-workforce.
14
Figure 6: Employment declines by worker characteristics
A. Feb−Apr Employment B. Feb−Aug Employment
Non US citizen Non US citizen
No college degree No college degree
Single Single
Born outside US Born outside US
Non−white Non−white
Female Female
Age < 50 Age < 50
Part time employed* Part time employed*
Below median wage* Below median wage*
Yes No Yes No
−0.30 −0.20 −0.10 0.00 −0.30 −0.20 −0.10 0.00
Feb−Apr log change in employment Feb−Aug log change in employment
Notes These figures plot employment changes by type of worker. Panel A plots employment changes from February
to April of 2020, adjusted by subtracting the average February-April change in employment for that group over 2010 to
2019. The variables on the y-axis are used to split workers into two groups: those in the group given by the label (‘Yes’,
marked with a green circle), and those outside that group (‘No’, marked with a red cross). For the whole sample, we
obtain a total decline in employment of −14.7 log points (black dashed line). Panel B is similarly constructed for
employment changes February-August 2020. For the whole sample, we obtain a total decline in employment of −7.1
log points (black dashed line) (∗) For the last two cases the sample is restricted to the Outgoing Rotation Group,
a subsample of the monthly CPS reporting hours—in the case of Part time employed —and hours+earnings—in the
case of Below median wage.
growth.23
4.2 Employment losses by worker characteristics
As a final exercise, we study how excess employment losses vary across the worker characteristics
considered in Section 3. For each group of workers we compute two objects: the total employment
change over February-April 2020, and the total employment change over February-August 2020. We
subtract off from each of these the mean total employment change across their respective months
for 2010-2019. We focus on employment rather than unemployment due to issues associated with
the labeling of workers as unemployed versus out of the labor force. Figure 6 shows the results.24
Once again, a clear pattern emerges: Figure 6A shows that those groups of individuals who have
a higher employment share in low WFH occupations (as identified using the methodology of Section
23
The metric we use to categorize occupations as essential is able to pick up certain obvious 2-digit occupations such
as healthcare technicians and healthcare support. However, some occupations are left out despite having numerous
mentions in the aforementioned government text. For example, the word construction is mentioned thirty three times;
the word legal is mentioned only once.
24
We check that the total employment losses February-April 2002 that we construct using survey weights lines up
with total employment losses reported by the BLS. We obtain a value of -14.7 percent for February-April. Because our
sample selection drops workers tenuously tied to the workforce, our estimate differs from the official value from the
BLS, -17.4 percent, which we compute as log( 133,403,000
158,759,000
) from Summary Table A of the following: BLS ‘Employment
Situation’ report - April, 2020.
15
2) experienced, on average, more severe employment outcomes in April 2020 relative to those in
occupations with high work-from-home capability.
For example, non-college workers (associated with in-person occupations) experienced a 21 per-
cent excess decline in employment from February to April 2020, while college educated workers
(associated with teleworkable occupations) saw only a 7 percent decline.
The characteristics with the largest differential in employment outcomes between groups are
income, work status, nativity, marital status, and education. Figure 6B depicts cumulative excess
employment losses from February through August. The data continue to show the same pattern as
in April, but the magnitudes of the differences in employment outcomes across groups decreased as
some in-person work returned.
4.3 Comparing employment losses to previous recessions
Our main results from Section 3 showed that economically more vulnerable workers were more likely
to be employed in highly exposed jobs as measured by our HPP and LWFH measures. However, since
these workers are generally more economically vulnerable, it may be the case that the employment
losses they suffered are independent from their exposure to the pandemic through their occupation,
and reflect standard employment dynamics in a recession.
To see whether individuals with the characteristics we have identified as being more exposed
suffered larger employment losses than would typically be expected, we conduct a simple difference-
in-difference exercise.
Table 1 compares peak to trough employment changes in LPP and HPP occupations during
both the Great recession and the Covid recession. We find that during the covid recession, high PP
occupations saw much larger employment losses relative to low PP occupations (−11.91 percentage
points). During the Great recession, High PP actually saw smaller decreases in employment relative
to low PP occuations (3.33 percentage points). This shows that high physical-proximity occupations
were uniquely exposed to the economic downturn induced by the novel Coronavirus and do not see
disproportionate employment losses during all recessions.
The outcome is similar but smaller in magnitude for our work-from-home classification. It is
true that in the Great Recession, workers in low work-from-home occupations saw 3.67 percentage
point larger employment losses than workers in high work-from-home occupations. However, in
the Covid recession, employment losses for workers in low work from home occupations were 9.14
percentage points larger than workers in high work-from-home occupations. This exercise confirms
that workers that cannot work-from-home or were in high physical proximity jobs were uniquely
exposed to the pandemic-induced recession.
16
Table 1: Employment losses by work-from-home and physical-proximity during two recessions
A. Work-from-home B. Physical proximity
Great Recession Covid Recession Great Recession Covid Recession
High WFH Low WFH High WFH Low WFH Low PP High PP Low PP High PP
A. Employment
Peak employment (level) 68.40 52.70 74.39 55.22 73.12 47.98 77.14 52.48
Trough employment (level) 66.27 49.12 67.64 45.16 68.70 46.69 70.85 41.95
B. Relative job losses
Percent change -3.12 -6.79 -9.08 -18.23 -6.04 -2.71 -8.16 -20.06
2nd difference -3.67 -9.14 3.33 -11.91
3rd difference -5.47 -15.24
Notes This table compares employment changes by LWFH and HPP during the Great Recession with the Covid
Recession. Panel A reports employment in millions at the peak and trough of both downturns. The peaks for the
Great and Covid recessions are January 2008 and February 2020. The troughs are January 2010 and April 2020.
Panel B reports percent decreases in employment for each occupation group from peak to trough. Second difference
compares percent decrease in employment between HWFH and LWFH and between LPP and HPP within a given
recession. Third difference compares relative employment losses between the Covid and Great Recession. LWFH
occupations saw 5.47 percentage point larger decrease in employment relative to HWFH occupations during the
Covid Recession than they did during the Great Recession. Similarly, HPP occupations saw a −15.24 percentage
point larger decrease in employment relative to LPP occupations during the Covid recession relative to the Great
Recession. LWFH and HPP occupations are not recession-exposed, rather there is a unique element of the Covid
recession which explains they larger drop in employment.
5 Conclusion
We show that workers systematically differ across the types of occupations that were most likely
to be hit by social distancing involved in both the public policy and behavioral responses to the
Coronavirus pandemic. Workers in occupations that are most likely to be affected—those with
a low score in the O∗ NET work-from-home measure, or a high score in the O∗ NET measure of
personal-proximity—are predominantly characterized by traits associated with the more econom-
ically vulnerable in the US economy. These workers are disproportionately less educated, have
limited healthcare, are toward the bottom of the income distribution, and have low levels of liquid
assets. We showed that this was a useful way of understanding job losses following the start of the
pandemic in 2020.
Given the occupation-level indicators we have constructed and made available with this paper,
our measures can be used to capture geographic or group level exposure to social distancing policies.
Moreover, our simple approach can be extended to individual economic indicators in any microdata
that records occupation, including, but not limited to, the Survey of Consumer Finances, the Survey
of Income and Program and Participation, and the Survey of Consumer Expectations.
17
References
Adams-Prassl, Abi, Teodora Boneva, Marta Golin, and Christopher Rauh, “Inequality
in the impact of the coronavirus shock: Evidence from real time surveys,” Journal of Public
Economics, 2020, 189 (C), S0047272720301092.
Akbarpour, Mohammad, Cody Cook, Aude Marzuoli, Simon Mongey, Abhishek Na-
garaj, Matteo Saccarola, Pietro Tebaldi, Shoshana Vasserman, and Hanbin Yang,
“Socioeconomic Network Heterogeneity and Pandemic Policy Response,” Working Paper 27374,
National Bureau of Economic Research June 2020.
Alexander, Diane and Ezra Karger, “Do Stay-at-Home Orders Cause People to Stay Home?
Effects of STay-at-Home Orders on Consumer Behavior,” Technical Report 2020.
Baqaee, David, Emmanuel Farhi, Michael J Mina, and James H Stock, “Reopening
Scenarios,” Working Paper 27244, National Bureau of Economic Research May 2020.
, , Michael Mina, and Jim Stock, “Evaluating the Economics and Epidemiology of Strate-
gies for Reopening the Economy,” 2020.
Bick, Alexander, Adam Blandin, and Karel Mertens, “Work from Home After the COVID-
19 Outbreak,” Working Papers 2017, Federal Reserve Bank of Dallas June 2020.
Chodorow-Reich, Gabriel, “The Employment Effects of Credit Market Disruptions: Firm-level
Evidence from the 2008-9 Financial Crisis,” The Quarterly Journal of Economics, 2014, 129 (1),
1–59.
Dingel, Jonathan and Brent Neiman, “How Many Jobs Can Be Done At Home?,” BFI White
Paper, Becker Friedman Institute 2020.
Goolsbee, Austan and Chad Syverson, “Fear, lockdown, and diversion: Comparing drivers of
pandemic economic decline 2020,” Journal of Public Economics, 2021, 193, 104311.
Gottlieb, Charles, Jan Grobovšek, Markus Poschke, and Fernando Saltiel, “Working
from Home in Developing Countries,” European Economic Review, 2021, p. 103679.
Heathcote, Jonathan, Fabrizio Perri, and Giovanni L. Violante, “Unequal We Stand: An
Empirical Analysis of Economic Inequality in the United States: 1967-2006,” Review of Economic
Dynamics, January 2010, 13 (1), 15–51.
Kaplan, Greg and Giovanni L. Violante, “A Model of the Consumption Response to Fiscal
Stimulus Payments,” Econometrica, July 2014, 82 (4), 1199–1239.
18
, Benjamin Moll, and Giovanni L Violante, “The Great Lockdown and the Big Stimulus:
Tracing the Pandemic Possibility Frontier for the U.S.,” Working Paper 27794, National Bureau
of Economic Research September 2020.
, Giovanni L. Violante, and Justin Weidner, “The Wealthy Hand-to-Mouth,” Brookings
Papers on Economic Activity, 2014, 45 (1 (Spring), 77–153.
Leibovici, Fernando, Ana Maria Santacrue, and Matthew Famiglietti, “Social Distancing
and Contact-Intensive Occupations,” Technical Report, St. Louis Federal Reserve Bank - On the
Economy Blog March 2020.
Lekfuangfu, Warn N., Suphanit Piyapromdee, Ponpoje Porapakkarm, and Nada Wasi,
“On Covid-19: New Implications of Job Task Requirements and Spouse’s Occupational Sorting,”
CReAM Discussion Paper Series 2012, Centre for Research and Analysis of Migration (CReAM),
Department of Economics, University College London April 2020.
Tomer, Adie and Joseph W. Kane, “How to protect essential workers during COVID-19,”
Brookings Report, Brookings Institute March 2020.
19
Appendix
A Data
In this appendix, we describe the construction of our work-from-home and personal-proximity mea-
sures in detail. We begin with an overview of the various datasets before detailing the construction
of our pandemic-exposure measures.
A.1 Overview of data sources
O∗ NET. The Occupational Information Network (O∗ NET) is a database sponsored by the U.S.
Department of Labor; it contains standardized measures on occupation skills, abilities, knowledge
requirements, work activities, and work contexts. Workers in more than 900 occupations rate their
job from 1 to 5 along many work activities and work contexts.25 The data is published using
the granular O∗ NET-SOC occupation classification system. O∗ NET-SOC is a refinement of the
Standard Occupational Classification (SOC) system developed by the Office of Management and
Budget (OMB).26 In this paper we use O∗ NET database 24 which makes use of the 2010 taxonomy
of the O∗ NET-SOC occupation codes. We aggregate the variables to the SOC level by taking means
weighted by the number of survey respondents in each O∗ NET-SOC code.
OES. The Occupational Employment Statistics (OES) provides wage and employment data by
occupation and industry. We use the 2018 snapshot which makes use of the 2010 taxonomy of the
SOC codes (SOC 2010). We use this data to employment-weight when aggregating O∗ NET data
from the SOC classification to the coarser OCC occupational classification system used by the CPS.
We use industry-occupation employment counts to develop our classification of essential occu-
pations. If at least 75% of an occupation’s employment is in industries deemed essential according
to the Department of Homeland Security, then we call that occupation ‘essential’.27
CPS. The Current Population Survey (CPS) is the official source of many U.S. government statis-
tics. We use this monthly survey to examine employment outcomes by occupation. Starting in May
2020, the CPS appended a series of questions related to the coronavirus pandemic to the monthly
questionaire. We use this module to examine teleworking behavior in Figure 3 and Table B4. Table
A1 contains summary statistics from selected months of the CPS.
25
Survey questions available here: https://www.onetcenter.org/dl_files/MS_Word/Work_Context.pdf. Human
resource experts rank occupation skills and abilities using both their own expertise and the survey responses.
26
For example, O∗ NET-SOC codes ‘33-2011.01’ Municipal Firefighters and ‘33-2011.02’ Forest Firefighters both
map to the SOC code ‘33-2011’ Firefighters.
27
Our essential industry data comes from Tomer and Kane (2020).
20
Table A1: Monthly CPS - Summary statistics
February April August
Mean SD Mean SD Mean SD
I. Continuous variables
Age 43.55 11.26 43.57 11.26 43.54 11.28
Hourly wage∗ 26.30 21.12 26.25 19.53 26.48 18.83
Usual weekly hours∗ 39.92 10.04 40.01 10.03 40.22 9.91
Years of education 13.99 2.72 13.98 2.73 13.99 2.73
II. Binary variables
College degree 0.38 0.38 0.38
Male 0.53 0.53 0.53
Born in US 0.80 0.80 0.80
Married 0.59 0.59 0.59
US citizen 0.91 0.91 0.91
White 0.79 0.79 0.79
Full-time employed 0.91 0.91 0.92
Age 50+ 0.34 0.34 0.34
N 585737 577490 571602
Notes This table reports summary statistics from the monthly CPS in the months of February, April, and
August. Panel A reports the mean and standard deviation of continuous variables. Panel B reports
the share of workers with the listed characteristic. Variables marked with an asterisk come from the Out-
going Rotation Group, a subsample of the monthly CPS which asks questions about wages and hours.
The March 2019 Annual Social and Economics Supplement (ASEC) asks detailed questions on
work and income. We use this March supplement to examine worker characteristics by occupation
in Figure 4, Figure B2, Figure B3, and Figure B4. Table A2 displays summary statistics.
PSID. The Panel Study of Income Dynamics (PSID) is a longitudinal survey of more than 18,000
individuals. We make use of the detailed data on income, expenditures, and wealth to study
household liquidity patterns by occupation. We follow the definition of hand-to-mouth used Kaplan
and Violante (2014). Summary statistics from the 2017 wave are shown in Table A3.
A.2 Construction of pandemic exposure measures
We construct two measures that measure an occupation’s exposure to social distancing during a
pandemic. We sign these measures in terms of expected negative economic impacts of the crisis: (i)
low work-from-home (LW F H), and (ii) high physical-proximity (HP P ). To create our measures,
we combine the SOC-level O∗ NET data on work activities with employment counts from the OES.
Merging these two data sources allows us to aggregate the work activities data to the coarser OCC
classification system used by the datasets which contain data on individual workers (ATUS, CPS,
and PSID).
We first detail the construction of HP P and then describe the construction of LW F H which
follows many of the same steps. Our procedure delivers two continuous variables LW F H j and
21
Table A2: CPS - ASEC 2019 March Supplement Summary Statistics
Mean Median SD
I. Continuous variables
Age 43.12 42.00 11.03
Hourly wage 29.90 22.50 33.40
Log annual hours 7.57 7.64 0.37
Years of education 14.14 14.00 2.77
II. Binary variables
College degree 0.41
Male 0.52
Has employer healthcare 0.94
Owns home 0.70
Born in US 0.79
Married 0.63
Big firm (500+ emp.) 0.51
US citizen 0.90
White 0.78
Full-time employed 0.89
No unemployment last year 0.67
Age 50+ 0.31
N 66786
Notes This table reports summary statistics from the Annual Social and Economic Supplement attached to the 2019
March CPS. Panel A reports the mean, median, and standard deviation of continuous variables. Panel B reports
the share of workers with the characteristic. Sample is restricted following the Sample C procedure described in
Heathcote et al. (2010).
22
Table A3: PSID Summary statistics
Mean SD
I. Continuous variables
Age 53.14 18.05
Family income from wages and salaries 52,546.74 82,730.07
Labor income incl. welfare 53,156.19 82,636.23
Usual weekly hours 41.81 13.56
Net worth 385,612.95 1,282,929.78
II. Binary variables
College degree 0.39
Male 0.68
Owns home 0.60
Born in US 0.91
Married 0.43
White 0.78
Full-time employed 0.92
Age 50+ 0.57
N 9155
Notes This table reports summary statistics from the 2017 wave of the Panel Study of Income Dynamics (PSID).
Panel A reports the mean and standard deviation of continuous variables. Panel B reports the share of workers
with the characteristic.
HP P j and two binary variables LW F H ∗j and HP P ∗j that can be mapped into the occupational
codes contained in the CPS, ATUS, and PSID. Let j ∈ {1, . . . , J} denote a 3-digit OCC-code
occupation, which is the measure available in worker-level data. Let l ∈ {1, . . . , L} denote the fine
SOC-code categorization of occupations in O∗ NET and OES.
Construction of physical-proximity measure. O∗ NET publishes data on the physical-
proximity, ml , at work for over 900 SOC occupations. The physical proximity measure takes on
values ml ∈ [1, 5].28
1. We start with ml , the O∗ NET measure of physical-proximity at the SOC level.
2. We aggregate ml to the OCC level, using the OES to compute an employment-weighted
mean HP P j for all SOC occupations l ∈ j. To map SOC code occupations into OCC code
occupations we start with a cross-walk obtained from US Census, which we then substantially
edit and verify.29
28
Workers that respond to the survey administered by O∗ NET choose one of: 1 = ‘I don’t work near other
people (beyond 100ft)’, 2 = ‘I work with others but not closely (e.g. private office)’, 3 = ‘Slightly close (e.g.
shared office)’, 4 = ‘Moderately close (at arm’s length)’, 5 = ‘Very close (near touching)’. Publicly available
O∗ NET data consists of an average of these responses. For additional information regarding this question, see
https://www.onetonline.org/find/descriptor/result/4.C.2.a.3
29
The basic cross-walk from Census is available here: https://www2.census.gov/programs-
surveys/demo/guidance/industry-occupation/2010-occ-codes-with-crosswalk-from-2002-2011.xls
23
X
HP P j = ωl ml
l∈j
nl
ωl = P
l0 ∈j nl0
3. We assign the dummy HP Pj∗ = 1 (high physical-proximity) if the occupation is above the
employment-weighted median of HP P j (=3.6).30
M in
4. We then re-scale HP P j to the interval [0, 1] by subtracting HP P j and dividing by
M ax M in
HP P j − HP P j .
Construction of work-from-home measure. The procedure to construct LW F H j and
LW F H ∗j is similar to the above. We differ from Dingel and Neiman (2020) in how we aggregate
skills, but use their set of O∗ NET job characteristics.
1. We take the following 17 measures of SOC-level occupation attributes in the O∗ NET data
from Dingel and Neiman (2020). We index them by k = 1, . . . , K. In the underlying job
characteristic data, each takes on a value mlk ∈ [1, 5], representing the average response of
workers to an underlying survey in which the options are {1, . . . , 5}:
- Work Activities module: Performing General Physical Activities; Handling and Moving Objects;
Controlling Machines and Processes; Operating Vehicles, Mechanized Devices, or Equipment; Performing
for or Working Directly with the Public; Inspecting Equipment, Structures, or Material; Repairing and
Maintaining Electronic Equipment; Repairing and Maintaining Mechanical Equipment.
- Work Contexts module: Electronic Mail Use;31 Outdoors, Exposed to Weather; Outdoors, Under
Cover; Deal With Physically Aggressive People; Wear Specialized Protective or Safety Equipment such
as Breathing Apparatus, Safety Harness, Full Protection Suits, or Radiation Protection; Wear Common
Protective or Safety Equipment such as Breathing Apparatus Safety Harness, Full Protection Suits, or
Radiation Protection; Spend Time Walking and Running; Exposed to Minor Burns, Cuts, Bites, or
Stings; Exposed to Disease or Infections.
Within each 3-digit OCC occupation j, we take the employment-weighted average of mlk
across SOC occupations l ∈ j. This gives a measure for each occupation-attribute pair:
P P
mjk = l∈j ωl mlk , where ωl = nl / l0 ∈j nl0 .
30
Since the cut-off value of HP P j is 3.6, then HP Pj∗ = 1 for occupations in which the average response to the
survey question is at least 4. Our high physical-proximity occupations therefore represent occupations for which the
average respondent said they worked at arm’s length or less away from others.
31
In the case of Electronic Mail Use, we reverse the values such that a high value implies that the occupation is
less suited to working from home.
24
2. We convert these into binary variables m∗jk ∈ {0, 1} based on whether mjk ≥ 3.5. Since
we employment-weighted when computing mjk then m∗jk = 1 if “The average respondent
to the question in the underlying O∗ NET survey answered at least 4, where an answer of 4
corresponds to performing the given activity ’once a week or more but not every day’.” 32
3. We then construct a single measure for each occupation LW F H j by taking the unweighted
mean of m∗jk : LW F H j = K −1 K ∗
P
k=1 mjk . In words, this gives the fraction of the K low
work-from-home measures mjk for which occupation j has a high score. We rescale this to
[0, 1] by subtracting the minimum value and dividing by the maximum minus the minimum
values.
4. We then assign the binary variable LW F H ∗j = 1 (low work-from-home) if occupation j is
above the employment-weighted median value of LW F H j (=2).
To recap, by construction HP P ∗j and LW F H ∗j are binary variables that equal 1 for the occupa-
tions that are most likely to be effected by the epidemic and ensuing policies. Half of employment
is in HP P ∗j = 1 jobs and half of employment is in LW F H ∗j = 1 jobs.
B Additional figures and tables
B.1 Worker characteristic results
Table B1 reports the share of workers in LW F H ∗ = 1 and LW F H ∗ = 0 (i.e. HW F H ∗ ) with the
worker characteristics of interest. As discussed in Section 2, βby is the difference. Table B2 similarly
reports worker characteristics for HP P ∗ = 1 and HP P ∗ = 0 (i.e. LP P ∗ ).
To examine the characteristics of workers who are employed in occupations which are both
LWFH and HPP we report the share of workers which the characteristics of interest in each of
the following bins ordered from least pandemic-exposed to most pandemic-exposed: (i) HWFH∗
and LPP∗ (ii) HWFH∗ and HPP∗ (iii) LWFH∗ and LPP∗ (iv) LWFH∗ and HPP∗ . In Table B3
the left-most column reports the share of workers in (i) HWFH∗ and LPP∗ occupations with the
relevant characteristic. The second column reports the difference between the share of workers in
(ii) HWFH∗ and HPP∗ relative to group (i). The third column reports the difference in the share
of workers in (iii) LWFH∗ and LPP∗ occupations with the characteristic relative to column (ii).
Column (iv) reports the difference between LWFH∗ and HPP∗ occupations compared to column
(iii).
Figure B2 extends the main results plotted in Figure 4 to examine disproportionate-ness condi-
tional on both HPP and LWFH.
25
Table A4: Occupation Rankings by LW F Hj and HP Pj
Occupation Employment share Rank LW F Hj Rank HP Pj
Install/Maintenance/Repair 0.04 1 10
Construction/Extraction 0.04 2 7
Healthcare supp. 0.03 3 1
Production 0.06 4 13
Material moving 0.03 5 11
Protection services 0.02 6 5
Building maintenance 0.03 7 16
Transport 0.04 8 12
Food prep. 0.09 9 4
Farm/Fish/Forest 0.003 10 22
Healthcare tech. 0.06 11 2
Personal care 0.04 12 3
Sales 0.10 13 8
Science 0.01 14 20
Community/Social 0.02 15 9
Office/Admin 0.15 16 15
Entertainment/Media 0.01 17 14
Management 0.05 18 19
Architecture/Engineering 0.02 19 18
Business/Financial 0.05 20 21
Computer/Math 0.03 21 17
Legal 0.01 22 23
Education 0.06 23 6
Notes This table ranks 2 digit OCC occupations by inability to work-from-home (LW F H ∗ ) and physical-proximity
(HP P ∗ ). Construction of LW F H ∗ and HP P ∗ described in Appendix A. Employment data is from OES.
Table B1: Share of workers in LWFH* and HWFH* occupations with the following characteristic
Worker characteristic Share of LWFH Share of HWFH Difference
No college degree 0.82 0.42 0.40
Below median wage 0.55 0.32 0.23
Male 0.63 0.45 0.18
Rents home 0.38 0.28 0.10
Born outside US 0.26 0.17 0.09
Single 0.46 0.37 0.09
Unemployed at all last year 0.38 0.30 0.08
Small firm (<500 emp.) 0.54 0.46 0.08
Non-US citizen 0.14 0.06 0.08
Non-white 0.25 0.21 0.03
Part-time employed 0.12 0.10 0.03
No employer healthcare 0.08 0.05 0.02
Age <50 0.67 0.67 0.01
Notes This table reports the share of workers in HW F H ∗ and LW F H ∗ occupations with the characteristics lists in
the leftmost column. The data comes from the 2019 March ASEC Supplement to the CPS.
26
Table B2: Share of workers in HPP* and LPP* occupations with the following characteristic
Worker characteristic Share of HPP Share of LPP Difference
Below median wage 0.53 0.34 0.19
No college degree 0.65 0.53 0.12
Part-time employed 0.16 0.07 0.09
Single 0.45 0.38 0.07
Rents home 0.35 0.30 0.05
Age <50 0.69 0.65 0.04
Non-white 0.25 0.21 0.03
Born outside US 0.22 0.20 0.03
Non-US citizen 0.11 0.09 0.01
No employer healthcare 0.07 0.06 0.01
Small firm (<500 emp.) 0.49 0.49 0.01
Unemployed at all last year 0.32 0.36 -0.04
Male 0.47 0.57 -0.10
Notes This table reports the share of workers in HP P ∗ and LP P ∗ occupations with the characteristics lists in the
leftmost column. The data comes from the 2019 March ASEC Supplement to the CPS.
Table B3: Share of workers in occupations with the following characteristic
Worker characteristic (i) HW F H ∗ & LP P ∗ (ii) HW F H ∗ & HP P ∗ (iii) LW F H ∗ & LP P ∗ (iv) LW F H ∗ & HP P ∗
No college degree 0.40 +0.05 +0.43 -0.11
Male 0.49 -0.15 +0.42 -0.22
Age < 50 0.66 +0.02 -0.04 +0.06
Small firm (< 500 emp.) 0.46 +0.00 +0.11 -0.05
Below median wage 0.26 +0.22 +0.07 +0.00
Single 0.36 +0.04 +0.04 +0.04
Unemployed at all last year 0.35 -0.12 +0.15 +0.00
Rents home 0.27 +0.02 +0.08 +0.01
Born outside US 0.17 +0.01 +0.10 -0.02
Non-white 0.21 +0.01 +0.00 +0.05
Non-US citizen 0.07 +0.00 +0.10 -0.03
No employer healthcare 0.05 +0.02 +0.02 -0.01
Part-time employed 0.07 +0.11 -0.10 +0.06
Notes Data is from the March 2019 ASEC supplement to the CPS. The column labeled (i) reports the share of
workers employed in HW F H ∗ & LP P ∗ occupations with the individual characteristics listed on the left-most
column. The columns labeled (ii) HW F H ∗ & HP P ∗ , (ii) LW F H ∗ & LP P ∗ , (iii) LW F H ∗ & HP P ∗ report the
difference in the share of workers with the worker characteristic relative to the column immediately to the left.
For example, 40% of workers in (i) HW F H ∗ & LP P ∗ occupations do not have a college degree. (ii) HW F H ∗ &
HP P ∗ have 40% + 5% = 45% of their workers reporting that they do not have a college degree. Occupations which
are (iii) LW F H ∗ & LP P ∗ have 40% + 5% + 43% = 88% of their workers reporting that they do not have a college
degree.
27
Table B4: Share teleworking due to Covid
Month All HW F H ∗ LW F H ∗ Difference
May 0.36 0.55 0.09 0.46
June 0.32 0.49 0.08 0.41
July 0.26 0.41 0.06 0.35
August 0.24 0.39 0.05 0.34
September 0.23 0.36 0.05 0.31
October 0.21 0.35 0.04 0.30
November 0.22 0.35 0.05 0.30
Notes This table reports the share of all workers and the share of workers in HWFH∗ and LWFH∗ occupations who
worked-from-home due to the coronavirus pandemic. Difference is the share of workers in HWFH∗ minus the share
of workers in LWFH∗ occupations who teleworked. Construction of HW F H ∗ and LW F H ∗ measures is described in
Appendix A. Survey responses come from the Covid module appended to the CPS beginning in May 2020.
B.2 Telework behavior during the Pandemic
B.3 Additional figures
(i) Professional, management, technology; (ii) Services (iii) Clerical, sales; (iv) Production, operators
1.0 Professional, management, tech. 1.0 Clerical, sales
Services Production, operators
Low work−from−home >> Low work−from−home >>
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0.0 0.0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
High physical−proximity >> High physical−proximity >>
Figure B1: Occupations by Work-from-home and Physical-proximity (3 digit, Census OCC)
Notes: This figure compares groups of 3 digit OCC code occupations. We split the data into panels A and B only for
readability, so that occupation titles can be included. The gray line plot fitted values and 95% confidence intervals
from an employment-weighted linear regression across all 3 digit occupations. Occupations above the red-dashed
line have LW F Hj∗ = 1, and account for half of employment. Occupations to the right of the blue-dashed line have
HP Pj∗ = 1, and account for half of employment.
As a robustness check, we confirm that our results are monotonic.
32
We choose a cutoff of for mjk of 3.5 exactly because we wanted to be able to make this kind of statement. If we
chose a cutoff of 4, then the average respondent would have answered above 4.
28
Table B5: Share teleworking due to Covid-19 Pandemic
(1) (2) (3) (4) (5)
LWFH Dummy -0.354∗∗∗ -0.329∗∗∗ -0.327∗∗∗ -0.321∗∗∗
(0.00197) (0.00200) (0.00198) (0.00224)
HPP Dummy -0.184∗∗∗ -0.0943∗∗∗ -0.0928∗∗∗ -0.0916∗∗∗
(0.00222) (0.00212) (0.00210) (0.00209)
May 0 0
(.) (.)
June -0.0434∗∗∗ -0.0434∗∗∗
(0.00445) (0.00445)
July -0.0935∗∗∗ -0.0937∗∗∗
(0.00433) (0.00433)
August -0.115∗∗∗ -0.116∗∗∗
(0.00421) (0.00421)
September -0.125∗∗∗ -0.125∗∗∗
(0.00407) (0.00407)
October -0.138∗∗∗ -0.138∗∗∗
(0.00399) (0.00399)
November -0.137∗∗∗ -0.137∗∗∗
(0.00402) (0.00402)
Essential -0.0184∗∗∗
(0.00280)
Constant 0.413∗∗∗ 0.337∗∗∗ 0.440∗∗∗ 0.534∗∗∗ 0.543∗∗∗
(0.00172) (0.00163) (0.00185) (0.00346) (0.00378)
Observations 192229 192229 192229 192229 192229
R2 0.159 0.042 0.169 0.181 0.182
Notes Standard errors in parentheses. Regressions use data from Covid module appended to the monthly CPS. We
regress the binary variable of whether a worker teleworked due to the coronavirus pandemic on our measures of an
occupation’s pandemic exposure. Occupations are essential if more than 75% of that occupations workers are
employed industries marked as essential by Tomer and Kane (2020). ∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001
29
A. Work−from−home B. Physical proximity
Below median wage No college degree
Poor hand−to−mouth* Below median wage
Hand−to−mouth* Male
Part−time employed Hand−to−mouth*
No college degree Poor hand−to−mouth*
Single Born outside US
Rents home Small firm (<500 emp.)
Age <50 Rents home
No employer healthcare Non−US citizen
Non−white Single
Born outside US No employer healthcare
Small firm (<500 emp.) Unemployed at all last year
Non−US citizen Part−time employed
Unemployed at all last year Non−white
High PP Low WFH
Male Low PP Age <50 High WFH
−0.20 0.00 0.20 0.40 0.60 −0.20 0.00 0.20 0.40 0.60
More likely >> More likely >>
Figure B2: Worker characteristics jointly examining physical-proximity and work-from-home
Notes This figure extends Figure 4 to jointly examine differences in worker characteristics of occupations using both
work-from-home and physical-proximity. This figure plots estimates of β y for 10 worker characteristics y. In Panel
A, the independent variable is LWFH∗ ∈ {0, 1}, our measure of inability to work-from-home. We run each regression
twice, conditioning on HPP∗ . Conditional on the occupation requiring low physical-proximity (red cross), a point
estimate β y = 0.25 shows that workers in both low PP and low WFH occupations are 25 ppt more likely to be below
median wage than occupations which are both low PP and high WFH. The blue dots condition similarly on high PP.
In Panel B, the independent variable is HPP∗ ∈ {0, 1}, our measure of high physical proximity required to perform
the job. We run each regression twice, conditioning on LWFH∗ . Blue dots compare HPP and LPP occupations which
are both low WFH. Red crosses compare HPP and LPP occupations which are both high WFH. We use O∗ NET
to construct our LWFH∗ and HPP∗ and use worker data from the CPS (non-asterisk) and PSID (asterisk) for the
regressions.
No college degree
Below median wage
Male
Born outside US
Single
Unemployed at all last year
Small firm (<500 emp.)
Non−US citizen
Non−white
Part−time employed
No employer healthcare Q2 vs Q3
Age <50 Q1+Q2 vs Q3+Q4
Q1 vs Q4
Rents home
−0.20 −0.10 0.00 0.10 0.20 0.30 0.40 0.50 0.60
Figure B3: Comparing different groups of occupations on the Work-from-home measure
Notes This figure extends Figure 4. The blue markers replicate Figure 4. In constructing the estimates plotted in
green, we set LW F Hj = 0 for the second quartile of our continuous measure z j , and LW F Hj = 1 for the third
quartile of z j . In constructing the estimates plotted in red, we set LW F Hj = 0 for the first quartile of z j , and
LW F Hj = 1 for the fourth quartile of z j .
30
Below median wage
No college degree
Part−time employed
Single
Age <50
Non−white
Born outside US
Non−US citizen
No employer healthcare
Small firm (<500 emp.)
Rents home Q2 vs Q3
Unemployed at all last year Q1+Q2 vs Q3+Q4
Q1 vs Q4
Male
−0.20 −0.10 0.00 0.10 0.20 0.30 0.40 0.50 0.60
Figure B4: Comparing different groups of occupations on the Physical Proximity measure
Notes This figure extends Figure 4. The blue markers replicate Figure 4. In constructing the estimates plotted in
green, we set HP Pj = 0 for the second quartile of our continuous measure z j , and HP Pj = 1 for the third quartile
of z j . In constructing the estimates plotted in red, we set HP Pj = 0 for the first quartile of z j , and HP Pj = 1 for
the fourth quartile of z j .
31
File and source
- File
- which-workers-bear-the-burden-of-social-distancing-policies.pdf
- Size
- 638,092 bytes
- SHA-256
- 10824914fa3bb4b2d2bde3a3560aa029e7d9a238c3285e9b366fec84a688495f
- Original
- www.nber.org