PreprintPDF Available
Preprints and early-stage research may not have been peer reviewed yet.

Abstract

Given the role of physical human proximity and contact in the spread of COVID-19, we build an index measuring the level of physical interaction for different occupations. Our Physical Interaction Index combines occupational work context information from O*NET and work travel information from the 2010 StatsSA Time Use Survey. We merge this with South African labour market data from 2018-2019 to explore the distribution of physical interaction across occupations and sectors shortly before the pandemic. The index provides some empirical evidence about a dimension of transmission risk that could inform how to calibrate the composition of economic sectors being phased back to work over the next few months. This short note introduces the index and provides some initial descriptive results for the South African labour market.
Jobs and COVID-19:
Measuring Work-Related
Physical Interaction
By Haroon Bhorat, Amy Thornton, Tim Köhler, and Morné Oosthuizen
DPRU Working Paper 202003
April 2020
JOBS AND COVID-19:
MEASURING WORK-RELATED PHYSICAL INTERACTION
DEVELOPMENT POLICY RESEARCH UNIT
HAROON BHORAT
AMY THORNTON
amy.thornton@uct.ac.za
TIM KÖHLER
MORNÉ OOSTHUIZEN
Working Paper 202003
ISBN 978-1-920633-72-1
April 2020
© DPRU, University of Cape Town 2020
This work is licenced under the Creative Commons Attribution-Non-Commercial-Share Alike 2.5 South Africa License. To view a copy of this lice nce,
visit http://creativecommons.org/licens es/by-nc-sa/2.5/za or send a letter to Creative Commons, 171 Second Street, Suite 300, San F rancisco,
California 94105, USA.
Abstract
Given the role of physical human proximity and contact in the spread of COVID-19, we build
an index measuring the level of physical interaction for different occupations. Our Physical
Interaction Index combines occupational work context information from O*NET and work
travel information from the 2010 StatsSA Time Use Survey. We merge this with South African
labour market data from 2018-2019 to explore the distribution of physical interaction across
occupations and sectors shortly before the pandemic. The index provides some empirical
evidence about a dimension of transmission risk that could inform how to calibrate the
composition of economic sectors being phased back to work over the next few months. This
short note introduces the index and provides some initial descriptive results for the South
African labour market.
Keywords
COVID-19; occupations; Physical Interaction Index; South Africa; work from home; work travel
Working Papers can be downloaded in PDF (Adobe Acrobat) format from www.dpru.uct.ac.za. A
limited number of printed copies are available from the Communications Manager: DPRU, University
of Cape Town, Private Bag X3, Rondebosch, Cape Town, 7700, South Africa. Tel: +27 (0)21 650 5701,
email: sarah.marriott@uct.ac.za
Corresponding author
Ms Amy Thornton
DPRU Researcher
c/o tel: +27 (0)21 650 5705
email: amy.thornton@uct.ac.za
Recommended citation
Bhorat, H., Thornton, A., hler, T. and Oosthuizen, M. (2020). Jobs and COVID-19: Measuring Work-
Related Physical Interaction. Development Policy Research Unit Working Paper 202003. DPRU,
University of Cape Town.
Disclaimer
The Working Paper series is intended to catalyse policy debate. They express the views of their
respective authors and not necessarily those of the Development Policy Research Unit (DPRU).
DPRU WP202003
1
1. Introduction
COVID-19 is a highly infectious virus that spreads from person to person when uninfected
people come into contact with respiratory droplets from an infected person who coughs or
sneezes (Naser et al., 2020). As such, the main tool policymakers have used to limit the spread
of the virus has been to lockdown their populations to prevent as much as possible
community transmission increasing exponentially and becoming unmanageable for the public
health system (Alvarez et al., 2020). In South Africa, schools and universities have closed and
a large portion of the employed have not been economically active since at least the 27th of
March 2020 when the national lockdown began. Kerr and Thornton (2020) estimate this
portion to be about two thirds of the employed, where the remaining third are either
classified as essential services or could work from home.
From the 1st of May 2020, the South African government will begin carefully easing the
lockdown according to areas where infection rates are lowest, and in industrial sectors where
workers have a lower transmission risk in the workplace (President Ramaphosa, 2020).
Evaluating which sectors put their workers at a higher or lower risk is an empirical question
which could be answered using data. There are many dimensions to answering this question,
some of which are epidemiological (infection rates per age), and others which are firm-
specific (how easily the work environment can be adapted to safety protocols) or relate to
the public health administration (the capacity of the public health system at a given point in
time). What we focus on in this document is another important dimension of work-related
transmission risk, and one about which we have data: the level of physical human interaction
people have on the job or on their way to the job. Our aim is to provide some evidence-based
insight into how physical human interaction was distributed across different sectors in South
Africa just before the pandemic hit.
To do this, we build an index of physical interaction for different occupations. Information on
physical interaction in the workplace comes from the Occupational Information Network
(O*NET), an American survey of detailed occupational information collected by the Bureau of
Labour Statistics. Examples include whether you share an office and how frequently you are
speaking to other workers face-to-face. Information on physical interaction in work travel
Jobs and COVID-19:
Measuring Work-Related Physical Interaction
2
comes from Statistics South Africa’s latest Time Use Survey. The scores per occupation are
then merged with South African labour market data in the Post-Apartheid Labour Market
Series (PALMS) version of the Quarterly Labour Force Surveys for 2018 and first two quarters
of 2019.
We are not the first to build an index of this type see Avdiu and Nayyar (2020) and Lu (2020)
and we draw on these efforts in our work. Our index does differ from these other examples.
We have tried to be clear about what we are, and are not, measuring. We are measuring
human physical interaction based on data from a pre-pandemic world of work. This is related
to, but is not the same thing as, transmission risk. As such, we have shied away from including
other aspects known to be related to COVID-19 transmission risk but not explicitly about
physical interaction, e.g. age distribution, occupations with contact with infectious diseases.
We have also tried to be stricter in our definition of physical interaction and so we exclude
the measures of team work and other contact included in other indices because these
measures can include non-physical contact like email or phone.
2. The Index
The Physical Interaction Index varies between zero and one and increases with the level of
physical interaction. There are three equally weighted dimensions: physical proximity (P),
face-to-face discussions (F), and use of public transport (T). The first two are drawn from
descriptions of occupational work contexts from O*NET, and the last is based on the 2010
Statistics South Africa Time Use Survey, based on our assumption that people who use public
transport to get to work have more physical interaction than those using private transport.
The definitions and scoring of each component are provided in Table 1. We impose explicit
equal weighting of components, following the lead of the Multidmensional Poverty Index
literature (Alkire & Foster, 2011). Explicit weighting keeps the index composition clear, and
we believe equal weights are justified in the case of our index. The three components are
combined as follows for occupation i at the four-digit level of occupation codes using the 2003
South African Standard Classification of Occupations (SASCO 2003):
Physical Interactioni = (⅓ * Pi) + (⅓ * Fi) + (⅓ * Ti)
DPRU WP202003
3
All components of the index are scaled to vary between zero and one. We first crosswalk the
O*NET components into ISCO-88 before merging with the South African labour market data
at the four-digit level (with adaption of Hardy’s (2016) resource). The component from the
StatsSA Time Use Survey naturally had compatible occupational codes and was directly
merged into PALMS.
So far, we make only one adjustment of the American O*NET data for a South African context.
Initial results scored domestic workers as one of the occupations with the lowest physical
interaction scores driven by a low physical proximity score. In South Africa, domestic workers
often perform a dual role of cleaning and child-minding leading us to think the physical
proximity score was too low for our context (du Plessis, 2018). We adjusted the proximity
score for domestic workers by replacing it with the mean of the physical proximity score for
domestic workers from O*NET and the physical proximity score for child-care workers (SASCO
code 5131). We believe this correction is justified given the importance and number of
domestic workers.
Table 1. Defining the components of the Physical Interaction Index
Component
Definition
Scoring
Physical
proximity
1. I don't work near other people
(beyond 100 ft.)
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)
O*NET spreads 100 points across
five levels per occupation. Our
approach multiplies points by their
category level and sums to get a
score. We sum points in categories
3-5 only to reach a score out of
500 (the maximum feasible score).
We rescale this to vary [0;1]
Face-to-face
discussions
1. Never
2. Once a year or more but not
every month
3. Once a month or more but not
every week
4. Once a week or more but not
every day
5. Every day
O*NET spreads 100 points across
five levels per occupation. Our
approach multiplies points by their
category level and sums to get a
score. We sum points in categories
4-5 only to reach a score out of
500 (the maximum feasible score).
We rescale this to vary [0;1]
Public
transport
Ever used any type of public
transport to travel to work on a given
day where public is defined as bus,
taxi, train and other transport and
private transport is defined as
walking, cycling, or private vehicle.
Share per occupation. Varies [0,1]
Use Survey,
2010
Jobs and COVID-19:
Measuring Work-Related Physical Interaction
4
3. Results: how does physical interaction vary across main
occupations?
We use a sample of the employed in the four quarters of 2018 and the first two quarters of
2019 to analyse our index. In the figure below, we collapse the index to the main occupational
code level. In this way we lose a lot of detail in the aggregation process, but this is still a useful
first exercise. As we would expect, people working in services have more physical interaction
than managers. We colour the bars with the contribution of each component. This allows us
to see that the use of public transport increases physical interaction for skilled agricultural
workers, whereas face-to-face discussions increases physical interaction for managers.
4. Results: how does physical interaction vary with the ability to work
from home?
Dingel and Neiman (2020) use O*NET to classify whether occupations can work from home
or not for the United States. Kerr and Thornton (2020) adapt this for the South African context
and also use the gazetted list of essential services to classify industries as essential or not at
the three-digit industry code level. In the figure below, we cross-reference our physical
DPRU WP202003
5
interaction index with their estimates for the ability to work from home for 25 sector
categories. The bubbles are weighted by employment share and coloured by the share of
essential workers in that sectoral category. We plot the data around the median for the
physical interaction index.
There is a negative correlation in the figure below meaning less physical interaction in the
workplace is associated with higher work-from-home potential. However, there is also a
cluster of sectors in the bottom left-hand corner where both physical interaction and the
ability to work from home are low. These sectors cover workers in agriculture; other
community, social, and personal services (many of the workers making up this bubble are
street sweepers); and slightly further up the physical interaction index, private households
including domestic workers.
The health sector has the highest score for physical interaction. It also has a very high share
of essential workers. Food trade and hotels and restaurants also rank highly in the physical
interaction index. By contrast, private households have the lowest score, and none of these
workers are classified as essential in the current lockdown (Alert Level 5). The finance sector
has a low level of physical interaction and the highest share of workers who could work from
home. This suggests that working from home would be a good strategy to keep transmission
risk low for this group. Manufacturing, the automotive trade sector, and non-food trade have
median levels of physical interaction, but very low shares of these sectors could work from
home.
Jobs and COVID-19:
Measuring Work-Related Physical Interaction
6
5. Conclusion
We believe our index is a useful method for providing some indication of where COVID-19
tranmission risk may be highest because of its allignment with physical interaction. As
mentioned previously, this index measures one aspect of transmission risk but is not an index
of transmission risk, itself. Other aspects may come into play. For example, it may be harder
to implement work safety protocols in a private household than a restaurant. These other
dimensions may ultimately reorder which occupations and sectors have a higher transmission
risk. Physical interaction though remains a key input into our understanding of how COVID-
19 spreads, and so we think providing some data on this topic may be helpful when choosing
the composition of sectors to phase back to work as the lockdown is eased. There are many
more potential applications this index could be used for, and interested researchers should
please contact the authors via amy.thornton@uct.ac.za.
DPRU WP202003
7
6. References
Alkire, S. and Foster, J. E. (2011). “Counting and multidimensional poverty measurement.” Journal of
Public Economics, 95(7): 476487.
Alvarez, F. E., Argente, D., and Lippi, F. (2020). A simple planning problem for covid-19 lockdown (No.
w26981). National Bureau of Economic Research.
Avdiu, B. and Nayyar, G. (2020) When face-to-face interactions become an occupational hazard: jobs
in the time of COVID-19. Brookings Future Development Article. Available:
https://www.brookings.edu/blog/future-development/2020/03/30/when-face-to-face-interactions-
become-an-occupational-hazard-jobs-in-the-time-of-covid-19/
Dingel, J. and Neiman, B. (2020). How Many Jobs Can be Done at Home? Available:
https://bfi.uchicago.edu/wp-content/uploads/BFI_White-Paper_Dingel_Neiman_3.2020.pdf
du Plessis, A. (2018). The role of domestic workers, as child carers, in the stimulation of motor
development of preschool children in Bloemfontein, South Africa. Master’s thesis. School of Allied
Health Sciences, University of the Free State.
Hardy, W. (2016). Occupation classification crosswalks from O*NET-SOC to ISCO. Institute for
Structural Research (IBS), Warsaw, Poland. Available online:
https://ibs.org.pl/en/resources/occupationclassifications-crosswalks-from-onet-soc-to-isco/
Kerr, A., Lam, D. and Wittenberg, M. (2019). Post-Apartheid Labour Market Series [dataset]. Version
3.3. Cape Town: DataFirst [producer and distributor], 2019.
Kerr, A. and Thornton, A. (2020). Essential workers, working from home and job loss vulnerability in
South Africa. A DataFirst Technical Paper 41. Cape Town: DataFirst, University of Cape Town.
Lu, M. (2020). The Front Line: Visualising the Occupations with the Highest COVID-19 Risk. Visual
Capitalist.com. Available: https://www.visualcapitalist.com/the-front-line-visualizing-the-
occupations-with-the-highest-covid-19-risk/
Naser, N., Masic, I., and Zildzic, M. (2020). Public Health Aspects of COVID-19 Infection with Focus on
Cardiovascular Diseases. Mater Sociomed, 32(1): 71-76
National Center for O*NET Development. O*NET OnLine. Retrieved April 24, 2020, from
https://www.onetonline.org/
President Ramaphosa, C. (2020) Televised address to the nation: 23 April 2020.
Statistics South Africa (2014). Time Use Survey 2010. [Dataset] Government of South Africa. Pretoria.
Statistics South Africa (2003). South African Standard Classification of Occupations (SASCO).
Available: http://www.statssa.gov.za/classifications/codelists/SASCO_2003.pdf
Development Policy Research Unit
University of Cape Town
Private Bag, Rondebosch 7701
Cape Town, South Africa
Tel: +27 21 650 5701
www.dpru.uct.ac.za
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Introduction: COVID-19 is the disease caused by an infection of the SARS-CoV-2 virus, first identified in the city of Wuhan, in China's Hubei province in December 2019. COVID-19 was previously known as 2019 Novel Coronavirus (2019-nCoV) respiratory disease before the World Health Organization (WHO) declared the official name as COVID-19 in February 2020. Aim: The aim of this study is to search scientific literature in the biomedicine and analyzed current results of investigations regarding morbidity and mortality rates as consequences of COVID-19 infection of Cardiovascular diseases (CVD), and other most common chronic diseases which are on the top mortality and morbidity rates in almost all countries in the world. Also, to propose most useful measures how to prevent patients to keep themselves against COVID-19 infection. Methods: We used method of descriptive analysis of the published papers with described studies about Corona virus connected with CVD, and, also, Guidelines proposed by World Health Organization (WHO) and European Society of Cardiology (ESC), and some other international associations which are included in global fighting against COVID-19 infection. Results: After searching current scientific literature we have acknowledged that not any Evidence Based Medicine (EBM) study in the world during last 5 months from the time when first cases of COVID-10 infection was detected. Also, there is no unique proposed ways of treatments and drugs to protect patients, especially people over 65 years old, who are very risk group to be affected with COVID-19. Expectations that vaccine against COVID-19 will be produced optimal during at least 10 months to 2 years, and in all current Guidelines most important proposed preventive measures are the same like which one described in Strategic documents of WHO, in statements of Declaration of Primary Health Care in Alma Ata in 1978. Conclusion: WHO proposed preventive measures can be helpful to everybody. Physicians who work at every level of Health Care Systems, but especially at primary health care level, must follow those recommendations and teach their patients about it. But, the fact is that current focus of COVID-19 epidemic has targeted on protection of physical health of population in global, however, the influence on mental health which will be one of the important consequences of COVID-19 pandemic in the future, and which could be declared as «Post-coronavirus Stress Syndrome" (PCSS) could be bigger challenge for Global Public Health.
Article
Full-text available
Multidimensional measures provide an alternative lens through which poverty may be viewed and understood. In recent work we have attempted to offer a practical approach to identifying the poor and measuring aggregate poverty (Alkire and Foster, J Public Econ, 2011). As this is quite a departure from traditional unidimensional and multidimensional poverty measurement—particularly with respect to the identification step—further elaboration may be warranted. In this paper we elucidate the strengths, limitations, and misunderstandings of multidimensional poverty measurement in order to clarify the debate and catalyse further research. We begin with general definitions of unidimensional and multidimensional methodologies for measuring poverty. We provide an intuitive description of our measurement approach, including a ‘dual cutoff’ identification step that views poverty as the state of being multiply deprived, and an aggregation step based on the traditional FGT measures. We briefly discuss five characteristics of our methodology that are easily overlooked or mistaken and conclude with some brief remarks on the way forward. KeywordsPoverty measurement–Multidimensional poverty–Deprivation–FGT measures–Decomposability–Joint distribution–Axiom
Article
The vulnerability of jobs to social distancing during the COVID-19 pandemic has thus far been analyzed through the lens of home-based work. However, as lockdowns ease, activities intensive in face-to-face interactions with consumers will likely be slower to recover as people continue to exercise social distancing precautions. This paper uses data on 900 job titles from the Occupational Information Network (O*NET) database for the United States to demonstrate substantial heterogeneity in the importance of face-to-face interactions across industries, income groups, and gender. First, industries vary in whether they emphasize face-to-face interactions and home-based work and the two do not always go hand-in-hand. Second, occupations that are more intensive in face-to-face interactions are largely concentrated among lower wage deciles. Third, a larger share of women’s employment is accounted for by occupations that are intensive in face-to-face interactions.
The role of domestic workers, as child carers, in the stimulation of motor development of preschool children in
  • J Dingel
  • B Neiman
Dingel, J. and Neiman, B. (2020). How Many Jobs Can be Done at Home? Available: https://bfi.uchicago.edu/wp-content/uploads/BFI_White-Paper_Dingel_Neiman_3.2020.pdf du Plessis, A. (2018). The role of domestic workers, as child carers, in the stimulation of motor development of preschool children in Bloemfontein, South Africa. Master's thesis. School of Allied Health Sciences, University of the Free State.
Occupation classification crosswalks -from O*NET-SOC to ISCO. Institute for Structural Research (IBS)
  • W Hardy
Hardy, W. (2016). Occupation classification crosswalks -from O*NET-SOC to ISCO. Institute for Structural Research (IBS), Warsaw, Poland. Available online: https://ibs.org.pl/en/resources/occupationclassifications-crosswalks-from-onet-soc-to-isco/
Post-Apartheid Labour Market Series
  • A Kerr
  • D Lam
  • M Wittenberg
Kerr, A., Lam, D. and Wittenberg, M. (2019). Post-Apartheid Labour Market Series [dataset]. Version 3.3. Cape Town: DataFirst [producer and distributor], 2019.
Essential workers, working from home and job loss vulnerability in South Africa. A DataFirst Technical Paper 41
  • A Kerr
  • A Thornton
Kerr, A. and Thornton, A. (2020). Essential workers, working from home and job loss vulnerability in South Africa. A DataFirst Technical Paper 41. Cape Town: DataFirst, University of Cape Town.
The Front Line: Visualising the Occupations with the Highest COVID-19 Risk. Visual Capitalist
  • M Lu
Lu, M. (2020). The Front Line: Visualising the Occupations with the Highest COVID-19 Risk. Visual Capitalist.com. Available: https://www.visualcapitalist.com/the-front-line-visualizing-theoccupations-with-the-highest-covid-19-risk/
South African Standard Classification of Occupations (SASCO)
Statistics South Africa (2003). South African Standard Classification of Occupations (SASCO). Available: http://www.statssa.gov.za/classifications/codelists/SASCO_2003.pdf