Article 2 - DISTANCING BONUS OR DOWNSCALING LOSS

DISTANCING BONUS OR DOWNSCALING LOSS?
THE CHANGING LIVELIHOOD OF US ONLINE

WORKERS IN TIMES OF COVID-19
FABIAN STEPHANY*, MICHAEL DUNN**, STEVEN SAWYER*** &
VILI LEHDONVIRTA*
* Oxford Internet Institute, University of Oxford, 1 St Giles, OX1 3JS, Oxford, UK. Email: fabian.
stephany@oii.ox.ac.uk; vili.lehdonvirta@oii.ox.ac.uk
** Department of Management & Business, Skidmore College, 815 North Broadway, Saratoga Springs, NY
12866, USA. Email: mdunn@skidmore.edu
*** School of Information Studies, Syracuse University, 343 Hinds Hall, Syracuse, NY 13244-1190, USA.
E-mail: ssawyer@syr.edu
Received: April 2020; accepted May 2020
ABSTRACT
We draw on data from the Online Labour Index and interviews with freelancers in the United
States securing work on online platforms, to illuminate effects of the COVID-19 pandemic. The
pandemic's global economic upheaval is shuttering shops and offices. Those able to do so are now
working remotely from their homes. They join workers who have always been working remotely:
freelancers who earn some or all of their income from projects secured via online labour platforms.
Data allow us to sketch a first picture of how the initial months of the COVID-19 pandemic have
affected the livelihoods of online freelancers. The data shows online labour demand falling rapidly
in early March 2020, but with an equally rapid recovery. We also find significant differences between
countries and occupations. Data from interviews make clear jobs are increasingly scarce even as
more people are creating profiles and seeking freelance work online.
Key words: Online labour markets; multi-method study; freelance work; COVID-19; telework
INTRODUCTION writing and editing, and other types of cognitive

work. The growth in online labour markets, and
We combine data from the Online Labour the project- or task-based structure of freelance
Index (OLI) and an interview-based panel study labour, provides a unique window into effects of
of freelance workers in the United States to pro- a pandemic on work. The OLI provides a global
vide insight into the changes in online labour perspective; the panel study provides for a more
markets relative to the ongoing pandemic arising localised set of insights.
from the global spread of the novel coronavirus, We pursue this work recognising that as the
COVID-19. We do so to contribute evidence and scope and breadth of the global COVID-19
insight to the ways in which a global pandemic pandemic continues to grow, the implications
appears in the localised context of one of the to workers and labour markets grow more pro-
economies most impacted. The freelance work- found. Even conservative estimates of a con-
ers that are the focus of this analysis are those tained global outbreak are showing significant
taking on projects and contracts for knowledge global macroeconomic impacts (McKibbin &
work: technology support, software develop- Fernando 2020). Similarly, models at the coun-
ment, bookkeeping, accounting, web content, try level are showing serious economic impacts
Tijdschrift voor Economische en Sociale Geografie – 2020, DOI:10.1111/tesg.12455, Vol. 111, No. 3, pp. 561–573.
© 2020 The Authors. Tijdschrift voor Economische en Sociale Geografie published by John Wiley & Sons Ltd on behalf of Royal
Dutch Geographical Society / Koninklijk Nederlands Aardrijkskundig
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and
reproduction in any medium, provided the original work is properly cited.
562 FABIAN STEPHANY ET AL.
(Atkeson 2020). Early research is also showing States, Germany, and South Korea. The data
distinct demographic and industry differences show distinctly different geographic patterns
in the impact of the pandemic (Wenham et al. between countries, with US data showing a
2020; Stephany et al. 2020). particularly acute drop in demand for work-
Building on the emerging COVID-19 related ers, with a simultaneously sharp increase in
research, we examine a specific subset of labour supply of available workers.
markets – the online labour market and its work- The data also make clear that not all oc-
ers – to understand the near-term impacts of the cupations in the United States are experienc-
pandemic. Online freelancers are in precarious ing the shocks similarly. Tech and software
work arrangements generally, and the pandemic development occupations show a significant
presents a particularly challenging scenario to increase in both online labour demand and
them for at least three reasons. First, the online number of registered profiles. Because of
nature of their work makes them susceptible the significant market shock evident in the
to greater competition (Dunn 2017). Second, United States and its notable occupational
freelance work is project-based: there is little to differences, we complement the quantitative
no commitment between employer and worker view with interviews with US-based online
beyond the specifics of the project’s contract freelance workers, helping us to understand
(Wood et al. 2019). Finally, in many countries the significance, implications and lived expe-
their status as independent contractors leaves riences of the market shock to online free-
online freelancers in vulnerable positions espe- lance workers.
cially during economic downturns. For example,
in the United States, benefits such as health care
are tied to formal, full-time employment and are BACKGROUND
not provided to freelancers. Indeed, in compar-
ison with many industrialised countries, there Online labour platforms are websites that me-
are relatively fewer labour and employment reg- diate between buyers and sellers of remotely
ulations governing non-standard work arrange- deliverable cognitive work (Horton 2010). The
ments in the United States (ILO 2016; McKay clients range from individuals and early-stage
et al. 2019). startups to Fortune 500 companies (Corporaal
On the one hand, online labour markets & Lehdonvirta 2017). The sellers are either
could be experiencing a boost in demand self-employed independent contractors, or
as companies move operations online. On people in regular employment who earn addi-
the other hand, the sharp economic down- tional income by moonlighting as freelancers
turn could be causing companies to reduce via the Internet. The platforms match clients
the use of online labour platforms alongside and workers using a variety of mechanisms,
other types of non-standard work. We address such as allowing clients to post projects for bid-
this issue by examining recent changes in the ding, and allowing freelancers to post resumes
global demand for online labour. And while for clients to evaluate. Besides matching, the
viruses may be blind to nation-states, policies platforms also handle contracting, time track-
and interventions are nation-state specific. ing, monitoring, billing, and dispute resolu-
Cross-national comparisons can therefore tion, allowing the entire relationship to be
provide insight on the economic implica- carried out remotely.
tions of specific policies and interventions. Online labour platforms can be further
Research has already begun to understand subdivided into freelancing platforms (e.g.
the economic implications of country specific Upwork, TopTal, Fiverr) where payment
interventions (e.g. Thunstrom et al. 2020). In is on an hourly or milestone basis, and mi-
this initial rapid analysis, we therefore use crotask platforms (e.g. Amazon Mechanical
quantitative data to examine changes in the Turk) where payment is on a piece rate basis
demand for online labour in three important (Lehdonvirta 2018). Of these, freelancing
regional economies with different counter- platforms appear to be much larger in terms
measures towards the pandemic: United of user numbers (Kässi & Lehdonvirta 2018).
DISTANCING BONUS OR DOWNSCALING LOSS? 563
Online labour platforms are also sometimes the pandemic, many skilled workers located
called online gig platforms, but they are in rural areas of the United States appear to
conceptually distinct from local gig econ- have successfully used online labour markets
omy platforms such as Uber or Deliveroo, to remotely access opportunities in urban areas
which involve physical on-site service deliv- (Braesemann et al. 2020). It is also possible that
ery (Wood et al. 2019). The global market some companies might be moving existing
for online labour has grown approximately contractor relationships to online labour plat-
50 per cent over the past three years (Kässi forms, in what is known as the ‘bring your own
& Lehdonvirta 2018). But as the COVID-19 freelancer’ model (Corporaal & Lehdonvirta
pandemic is hitting the world’s economies, 2017). Platforms provide features for monitor-
causing a massive rise in unemployment in ing and managing contractors remotely, which
the United States, it is pertinent to ask how can partly substitute for the missing in-per-
the pandemic is affecting the market for on- son controls (Braesemann et al. 2020). Short-
line labour. term contractors are typically subject to more
performance- and outcome-based controls
Pandemic’s potential effects on online labour than regular employees, whose loyalty may be
demand – There are several potential sought with cultural and incentive-based con-
mechanisms through which the pandemic trols (Ouchi 1979).
could be causing a positive demand shock Furthermore, the pandemic and the switch
for online labour. The pandemic appears to to remote work and virtual collaboration might
be forcing companies in affected countries also be creating additional demand for cer-
to shift from collocated office work towards tain types of labour, some of which is supplied
home-based remote working arrangements, through online labour platforms. In particular,
known in previous literature as telework (Huws it is conceivable that there is a significant posi-
et al. 1990) or telecommuting (Mokhtarian, tive demand shock for information technology
1991), and now colloquially referred to as contractors who can help companies set up
‘work from home’. The pandemic is also and maintain remote work and virtual collab-
likely pushing companies to increase the oration infrastructure, and provide training on
use of virtual collaboration as a substitute to their use. Systems integration work and data-
travel and face-to-face meetings. Telework base management work might also increase as
and virtual collaboration have been slowly organisational reliance on systems over in-per-
and unevenly gaining in popularity already son interactions increases. Large firms often
since their introduction in the early 1990s. have existing IT services outsourcing provid-
Now, there appears to be an unprecedented ers, but small- and medium-sized companies
surge of interest towards them. For instance, may be turning to online labour platforms for
interest in remote working and related search these needs (Bunyaratavej et al. 2011).
terms in Google search approximately tripled However, it is also conceivable that the pan-
from its pre-pandemic baseline to March demic could be causing a significant reduc-
2020 (Clement 2020). The stock market tion in demand for online labour, through
value of teleconferencing software company a few possible mechanisms. Online labour is
Zoom Video Communications approximately part of the broader category of non-standard
doubled in the same time period.1 work, which includes other independent con-
Given this surge of interest in remote work tractors and temporary workers. Many com-
and virtual collabouration, it is conceivable panies in the United States use non-standard
that remote-by-design online labour markets workers as a flexible buffer that can be rap-
could be seeing a significant demand boost. idly reduced in economic downturns to pro-
Companies looking to engage new contrac- tect core workers (Kalleberg 2003). Online
tors might now favour remote online con- labour can also be seen as a form of outsourc-
tractors hired through web-based platforms ing (Lehdonvirta et al. 2019), which is like-
over on-site contractors hired through con- wise something that companies can adjust
ventional staffing agencies. Already before to respond to changing demand. Given that
the pandemic and its public policy responses loyalty to their employers. Both moonlighters
have in many countries already resulted in a as well as existing full-time online freelancers
general economic downturn of historic pro- might also decrease the hours supplied online
portions, it is conceivable that companies as a result of falling ill or having increased care
could be cutting their use of online labour and housework duties, as schools and daycare
platforms to engage contractors. Already be- centres close or family members fall ill.
fore the pandemic it was apparent that de-
mand in the online labour market was very Geographic and temporal variation in the
responsive to events such as public holidays effects – The pandemic’s possible positive and
(Kässi & Lehdonvirta 2018), and workers felt negative effects on online labour supply and
that demand for their services could fluctuate demand are likely to vary across space and time,
a lot (Lehdonvirta 2018). for a variety of reasons. Some reasons have to
Two opposing effects of the COVID-19 pan- do with the dynamics of the pandemic itself.
demic on demand for online freelance labour The pandemic unfolds in phases, from initial
are thus conceivable. On the one hand, to the discovery to growing spread and eventual
extent that online labour is a substitute to on- public and private countermeasures, followed
site labour, we can expect to see an increase by an easing up of the countermeasures and
in demand. The technical and organisational possible additional waves of infection (Bedford
effort involved in switching to remote oper- et al. 2020). Owing to differences in initial entry
ations might also cause an increase in the time, infection rates, and response speed,
demand for specific types of online labour, es- different countries are at different phases,
pecially IT services. We will refer to all these with different impacts on economic activity.
demand-increasing mechanisms collectively as Country differences in public and private
the distancing bonus. On the other hand, to the countermeasures, severity of the epidemic, and
extent that online labour is a complement to underlying economic conditions are also likely
general economic activity, we can expect to see to generate geographic variation in effects
a decrease in demand, as companies facing de- seen in online labour markets (ILO 2020).
clining revenues reduce non-essential spend- Moreover, dynamics of the online labour
ing, including external online contractors. We market itself provide plenty of reasons to ex-
refer to this as the downscaling loss. pect geographic differences (Beerepoot &
Lambregts 2015). Generally speaking, a clear
Pandemic’s potential effects on online labour Global North--Global South trade pattern is ev-
supply – The pandemic can also be expected ident in online labour markets, where the ma-
to have a significant impact on the supply of jority of employers are located in high-income
labour on online labour markets. The supply countries (Kässi & Lehdonvirta 2018) while
generally speaking consists of two margins: most workers are located in low- and middle-in-
the number of workers offering their services come countries (Lehdonvirta et al. 2019).
through online platforms, and the number of Different employer countries have different
hours that they are willing to supply (Horton demand profiles in terms of the occupations
2010). The number of workers offering services they are buying from online labour platforms,
online might conceivably increase, because the although these differences are surprisingly
pandemic and its countermeasures have led to small (Kässi & Lehdonvirta 2018).
record-high unemployment in many countries, Changes in supply and demand in different
freeing up skilled workers (del Rio-Chanona countries are ultimately reflected in the global
et al. 2020). Workers who were already offering aggregate online labour supply and demand.
a small number of hours online on top of their Although there is evidence of preference
regular work (i.e. moonlighting; Pesole et al. and discrimination in online labour markets
2018) might increase the hours offered as a (Galperin & Greppi 2019), overall the market
result of being laid off. On the other hand, functions globally, with workers with similar
workers who are not laid off from their regular skills and experience considered reasonably
jobs might decrease the number of hours they good substitutes for each other regardless of
supply through online labour markets, to signal location (Lehdonvirta et al. 2019).
The resulting overall effects on the market an index that measures the utilisation of online
have substantive policy implications. If and labour platforms over time and across coun-
when online labour demand experiences a tries and occupations (Kässi & Lehdonvirta
distancing bonus effect while supply remains 2018). It serves a similar function as conven-
relatively constant, then the growing online la- tional labour market statistics on new vacan-
bour market could offer an avenue for laid-off cies. The index is constructed by continuously
workers to recover some of their earnings. But collecting data on tasks and projects posted
if the downscaling loss effect dominates and/ on major online labour platforms in near re-
or the market is flooded with additional labour al-time. The results of the OLI are published
supply, then people thus far earning their main as an open data set and an interactive online
income through online labour platforms are visualisation, updated daily (http://ilabour.oii.
likely to join the ranks of self-employed peo- ox.ac.uk/online-labour-index/). In this rapid
ple urgently in need of financial assistance. In initial assessment, we examine the effects of
the following section we present a rapid initial the pandemic on the aggregate global market,
quantitative assessment of which effects domi- followed by three important regional econo-
nate, across time and selected countries. mies. Finally, we present additional analyses
pertaining to the United States.
QUANTITATIVE EVIDENCE FROM THE As discussed in the literature review, online
ONLINE LABOUR INDEX labour markets have a distinct geography. Our
data allows us to provide a detailed picture of
We draw on the OLI for a first quantitative as- how the demand for online labour is geograph-
sessment of how online labour markets are af- ically distributed on the platforms monitored
fected by the COVID-19 pandemic. The OLI is by the OLI. As shown in Figure 1, the largest
Figure 1. Globally, in 2020, the largest share of online labour demand stems from the United States. [Colour figure can
be viewed at wileyonlinelibrary.com]
share of online labour demand originates the other hand, indicates that at this stage of
from employers based in the United States, the pandemic, the distancing bonus may have
who posted 41 per cent of all projects recorded started to dominate over the downscaling loss.
in 2020. The second largest buyer country for To further examine this fast-changing dy-
online freelance work is the United Kingdom namic, we are interested in examining the
(8%), followed by India (6%). Europe exclud- biggest demander country, the United States.
ing the United Kingdom generates 16 per cent However, the United States was not the first
of online labour demand, with Germany as the economy to be impacted by the pandemic and
biggest demander in this group. Only three per its countermeasures. It is useful to contrast
cent of the demand for online workers comes it with other important regional economies
from the entirety of Africa. Given these very with different COVID-19 trajectories and re-
uneven geographies of online labour demand, sponses. In this study, we limit ourselves to ex-
it is fair to assume that global developments on amining three important regional economies:
online labour markets are often driven by buy- Germany, South Korea, and the United States
ers from the United States. as the three economies confronted the pan-
demic at different points in time and opted for
Changes in online labour demand – Over the different countermeasures.
past several years, the OLI has shown a clear Figure 3 shows a cross-country comparison
seasonal pattern: demand drops during the of online labour demand from Germany, South
year-end holiday season, and then rises again Korea and the United States. Each country
to reach a plateau in February, which normally shows a distinct pattern. South Korea was one
persists until May. However, as Figure 2 shows, of the first economies that had to face the con-
this is not the case in 2020. By mid-March, sequences of COVID-19 early in 2020. Demand
when the World Health Organisation declared from South Korea fell from mid-February to
that COVID-19 had become a pandemic, the mid-March, but bounced back rapidly in late
OLI was in deep decline, in comparison to March. Demand from Germany similarly fell
2018 and 2019. This finding indicates that the from February to March, and experienced a
downscaling loss effect may be dominating more modest rebound in April. Demand from
over the distancing bonus effect. However, in the United States started falling roughly two
early April 2020 the OLI began to rise again, weeks later than the demand from South Korea
surpassing by far the usual level of previous and Germany and fell furthest, but by early May
years by the end of April. This observation, on was close to reaching pre-crisis levels again.
Figure 2. Demand on major online labour platforms, 1 January to 18 April (2018–2020), relative to the start of the year
[Colour figure can be viewed at wileyonlinelibrary.com]
Figure 3. Online labour demand from the United States, Germany and South Korea, relative to the start of the year.
[Colour figure can be viewed at wileyonlinelibrary.com]
These patterns are consistent with the idea of these categories, see Kässi and Lehdonvirta
that the pandemic reached the United States (2018).
later than it reached South Korea and Germany. As Figure 4 shows, not all occupations have
Moreover, in comparison to the United States, experienced a drop in demand. Demand in
Germany and South Korea imposed counter- creative and multimedia or sales and market-
measures of different types and magnitudes. ing support has shrunk significantly as the pan-
With the historical experience of the MERS demic has unfolded. But requests for projects in
outbreak in 2015, South Korea relied on mass the software development and technology cat-
testing and digital infection tracing, allowing egory remain largely unaffected. This finding
the country to avoid a full lockdown (Normile is consistent with an interpretation that com-
2020). Germany established early and localised panies are cutting non-essential freelance con-
testing, which allowed the government to im- tracts, such as marketing and sales campaigns,
pose only relatively moderate restrictions and while maintaining freelance outsourcing that
to permit local business to open again in the is essential for continued business operations,
beginning of April. Demand from the United such as tech support and database management.
States can also be seen bouncing back from The figures are also consistent with the idea that
mid-April onwards, even though restrictions the rapid push towards videoconferencing and
remained largely in place. This could reflect other remote operations across companies has
businesses adjusting to the new normal of re- created additional demand for freelance IT spe-
mote work, rather than a return to pre-pan- cialists who are able to help with this.
demic operations.
Given the particularly clear drop in de- Changes in online labour supply – The OLI
mand for online labour from the United does not provide us with a direct measure of
States, we are interested in how this drop labour supply, but we are able to observe the
may vary across different types of work. The number of registered worker profiles on a
OLI categorises online labour into six dif- smaller set of online labour platforms.1 This
ferent occupations: clerical and data entry, can be used as an imperfect proxy for the
professional services, software development number of workers offering services through
and technology, creative and multimedia, online labour platforms. We are not able to
sales and marketing support, and writing and observe changes in the number of hours the
translation. For a more detailed description workers are supplying.
Figure 4. Online labour demand from the US, relative to the start of the year, by occupation [Colour figure can be viewed
at wileyonlinelibrary.com]
An increase in the number of registered workers’ experiences, we draw on data from

freelancer profiles in the United States is ev- an ongoing panel study in the United States
ident since the beginning of April. In partic- that relies on structured interviews to pro-
ular, as shown in Figure 5, a very significant vide insights on how these trends are being
number of new freelancers have registered manifested through the lived experiences of
in the software development and technology workers.
category. Other occupations do not show a
similar increase in registered workers. This
is consistent with an interpretation that re- INSIGHTS FROM INTERVIEWS WITH
cently laid-off workers across the economy FREELANCE WORKERS IN THE UNITED
are not registering en masse on online labour STATES
platforms to attempt to make up for lost in-
come. Some workers are probably doing so, Our interview data come from an ongoing
but to some extent the increased supply in panel study of 60 freelance workers who are lo-
the software development and technology cated in the United States and seeking work on-
category might also be attributable to the line via the online labour platform Upwork (See
pull of increased opportunities due to grow- http://upwork.com). Upwork is one of many
ing demand. online labour platforms and routinely seen as
In sum, while we observe a downscaling a dominant player. The focus on Upworkers
loss for most online labour occupations in serves as a window into the career plans and
the United States, software and tech jobs ap- work strategies of freelance workers seeing work
pear to profit from a distancing bonus effect. online. The study is designed around a carefully
However, even in this category labour supply constructed sample of people who pursue free-
growth in the United States appears to have lance work as a primary or secondary source of
outpaced labour demand growth, suggesting income, and reflect a range of work types, skill
that the workers are likely to be experienc- levels, experience online, gender, ethnicity and
ing a tight market. To better understand the success with this work.
Figure 5. Registered online worker user accounts in the US, relative to the start of the year. [Colour figure can be viewed
at wileyonlinelibrary.com]
Participants are hired and paid as they design and this initial analysis, relies on the job
would for other jobs found on Upwork.2 Once classifications provided by Upwork,4 grouped
hired, participants complete a 15’ survey into three broader categories (See Table 1 for
that provides us an overview of their working summary statistics):
plans, outcomes and experiences and a 45’
interview. The interview builds on the survey 1. Administrative work to include: accounting,
data and follows a carefully designed protocol customer service, translation, editing. In
of semi-structured questions. Interviews are the OLI, this category roughly responds to
done by one of the six members of the digital the occupations ‘writing and translation’,
work research group, a joint effort of Syracuse ‘professional services’, and ‘clerical and
University and Skidmore College, both in New data entry’.
York state. The research team members were 2. Technology work to include: web, mobile
trained on the protocol and meet frequently and standard programming, engineering
and routinely to review the protocol, the data, and architecture, data and analytics, IT
and pursue interim analyses (as is customary in and networking. In the OLI, this category
field studies).
roughly responds to the occupation ‘soft-
Beginning in mid-March 2020, we asked
ware development and tech’.
freelancers how they were faring in the face of
3. Creative work to include: design, graphics,
the COVID-19 pandemic. Since then, we have
done 31 interviews. For this paper, we com- sales and marketing, writing, and some one-
pleted an interim analysis of these freelanc- offs like crisis response public-relations. In
ers, reviewing the transcripts of the interview, the OLI, this category roughly responds to
drawing on the field notes, and looking to sec- the occupations ‘creative and multimedia’,
ondary sources for additional insight.3 Specific ‘writing and translation’, and ‘sales and
to comparing with the OLI, the panel study marketing support’.
Table 1. Summary statistics.
Job classification Gender
Administrative 32% Female 61%

Technology 32% Male 39%
Creative 36%
Hours a week spent looking for work
Role of Freelance Work Min 0

Primary Source of Additional Source of Employment 1st Qu. 1
Employment
53% 47% Median 5
Mean 11
3rd Qu. 20
Max 45
Age Length on digital platforms (in years)
Min 23 Min 0
1st Qu. 30 1st Qu. 2
Median 35 Median 3
Mean 37.99 Mean 5.25
3rd Qu. 45 3rd Qu. 8.50
Max 75 Max 20
Data from the panel study provide sub- different type of work, increasing the num-
stantial evidence in support of downscaling ber of jobs they simultaneously are bidding
loss effects. The freelancers with whom we for, bidding for lower paying contracts and
spoke report that there are fewer jobs being lowering their hourly rates.
posted. Freelancers who have long-standing From this initial analysis we see little ev-
clients tell us these clients are pausing cur- idence of distancing bonus effects in the
rent projects and not adding new projects. United States across the platforms broadly,
Furthermore, freelancers indicate that there although we see initial evidence of distancing
are many more people bidding for the jobs: bonus effects at the occupation level. At first
sometimes six times more than even a month these changes were not affecting one broad
ago (e.g. 30+ bids now, versus five bids). category of work more than another. By April
This seems to align very well with the data 2020, those respondents in creative spaces
presented in Figure 5. Freelancers are also (marketing, design, web content) noted work
reporting that in response to the increased was slowing dramatically, while work support-
number of people seeking work online, they ing digital infrastructure (web services, back-
are having to bid on more jobs to keep se- end web work) was steady or even increasing,
curing work, and often are bidding for work corroborating with the data in Figure 4. This
below their target salary rates, and even below suggests that clients are starting to look into
minimum wage rates. While each freelancer’s the concerning future and prioritising basic
situation is unique, work-seeking strategies operations over customer-facing efforts.
in our initial analysis, seem to be externally Beyond these insights on distancing bonus
focused. That is, while many have relied on and downscaling loss effects, two additional
existing relationships, their predictability of observations from our interviews warrant men-
these relationships producing work are now tion. First, these US-based freelancers are re-
more tenuous. In turn, freelancers indicated porting changes in their life worlds that are
that they are diversifying and applying to reshaping their labour strategies and working
arrangements. Nearly all of the respondents re- access to social policies that could serve as a
port spouses and partners being laid off or hav- social safety net (e.g. access to healthcare and
ing to work from home, and that their children unemployment benefits). The broad and cha-
are now at home for schooling. These changes otic response by the United States’ Federal
are requiring the freelancers to alter their own Government6 in the face of the economic crisis
working arrangements and work availability – is more poignant for freelance workers as their
with most having less time to pursue work. The status relative to what counts as work often
scope of these changes is increasing each week, means they are not able to access unemploy-
and the changes are becoming more impact- ment benefits and may not be eligible for the
ful as the economy continues to slow down. stimulus checks associated with the CARES act.
The current arrangements are no longer new, Given the social and labour policies, significant
not yet normal, and uncertainty clouds things. market shocks, like the pandemic, make visible
These changes dampen the ability of workers the risks associated with non-standard work ar-
to take on typical work loads, providing some rangements in the United States.
opportunity for new entrants. These risks may be one of the reasons why
Finally, every one of the participants in the the OLI figures for the United States vary so
panel study is casting an increasingly worried eye much more than do those of South Korea and
on the longer term as the economic slowdown Germany. While this analysis has focused on
in the United States continues to press on most the United States, to fully understand the nu-
but the wealthy. Panel data make clear that fewer ances and differences in the online labour mar-
than 40 per cent of our respondents have health ket, future analysis exploring other countries is
care: a profoundly disturbing finding given that essential. Furthermore, the data also only pro-
many of their spouses and partners are losing vide a freelancer perspective, so understanding
jobs. Panel data also make clear that fewer than changes and behaviors of clients (employers)
40 per cent of our respondents had more than and their impact (and how they’re impacted)
a few weeks of savings: seven weeks into massive would provide an important perspective in the
social distancing and 22 million people losing market dynamics discussed here.
their jobs, these workers are exhausting these We see value in the further interrogation of
reserves, well before the economic crisis is over. the downscaling loss and distancing bonus effects
Taken together, the low number of freelancers from other analytic perspectives to include tem-
with healthcare, and savings, showcases the pre- porality, geography, and demography. The mag-
carity of the workspace and amplifies the im- nitude of downscaling and distancing are likely
plications of the changes to the online labour to change over time, as the pandemic and its
markets as demonstrated above. consequences unfold. Moreover the effects are
likely to vary across geography, due to national
differences in economic structures, the phase
DISCUSSION AND CONCLUSION and severity of the pandemic, and countermea-
sures adopted. We also see distinct differences in
Findings suggest the downscaling loss effect demands for different occupations. Considering
may be dominating over the distancing bonus the gendered and racial differences in occupa-
effect. At the same time, the number of work- tions (e.g. women, for example, are overrepre-
ers seeking income through online labour sented in the sales and creative occupations and
platforms is increasing. This, in turn, suggests minorities are overrepresented in service occu-
freelancers earning their income through on- pations, See https://conte xts.org/blog/inequ
line labour platforms are more likely to see a ality-during-the-coronavirus-pandemic/#rea),
tighter market and larger variations in their we then may see certain groups disproportion-
income in the face of a more uncertain online ately affected. As can be seen, one possible
labour market. For the US-based freelancers roadmap for near-term research related to the
with whom we spoke, their situation is further pandemic can be: (i) the examination in the
complicated by the structural and legal land- differences driven by occupation; (ii) analysis of
scape they face. To wit, their employment classi- the differences in country-level trajectories; and
fication (as independent contractors) restricts (iii) further investigation of gender and racial
disparities found in freelance work. These im- Notes

plications illustrate the questions that emerge
1. See https://markets.businessinsider.com/news/
from this early view into changes in the online
stock s /zooms t ock-price-surged-coron a virus-
labour markets for freelance work arising from
pandemic-video-work-from-home-2020-3-10290
the global pandemic.
23594
At a broader level, the COVID-19 pandemic
2. The raw data (used in Figure 5) are collected by
produces a unique window of opportunity to
periodically sampling workers from four major
gauge the interdependencies between the on-
online labour platforms: Fiverr, Freelancer,
line labour market and the regular economy.
Guru, and PeoplePerHour – http://ilabour.oii.
Initially, the sharp decline in the online labour
ox.ac.uk/measuring-the-supply-of-digital-labou
market closely mirrored the broader labour
r-how-the-oli-worker-supplement-is-constructed
market. As the pandemic has continued, our
3. To reduce the pressure on having Upworkers
analysis shows a rebound in the online labour
bidding for work (which consumes precious
market. This not only suggests that online la-
bidding resources), they are hired directly:
bour markets can transcend its traditionally
reaching out to invite them to take the work.
complementary role in the broader markets,
About 6 of 10 do: acceptances and declinations
but may play a key role in bridging gaps in the
are tracked. Research guidelines require paying
broader economy caused by sudden economic
those who agree to this job even if they do not
shocks. We speculate this is being manifested
finish: this work is voluntary and they can cease
in at least three ways. First, firms are using
working at any time. About 1 in 10 begin, but do
more project-based work, which is more con-
not finish. All who begin are rated five stars (as
ducive to the use of freelancers. Second, firms
performance ratings matter greatly for online
are increasingly becoming more accustomed
workers). Many of those who have completed
to distributed and remote working arrange-
the work leave a positive review of the work.
ments, which is also more conducive to free-
There are no negative reviews.
lancers. While both of these are independent
4. For example, there are active sub-threads for
of the current pandemic, they are macro-level
experiences with platform work, generally, and
changes in employment relations that we see
Upwork, specifically, to be found in the online
as catalysts to the increases in the demand for
platform Reddit, as well as Facebook groups de-
online freelancers. Third, online platforms
voted to both.
allow people to more easily seek freelance
5. Upwork changes its categorisation of workers from
work, which blurs the boundaries between on-
time to time, making classification a challenge.
line and traditional labour markets. As a more
6. CARES Act, see: https://www.congr ess.gov/
efficient matching system for work, online la-
bill/116th-congr e ss/senate-bill/3548/tex-
bour markets help to alleviate the friction as-
t?q=product+update.
sociated with job searches and allow workers
to seek work beyond their local geography. All
of these things allow online labour markets to REFERENCES
normalise more quickly during periods of eco-
nomic shock. As the evidence in our analysis Atkeson, A. (2020), What Will Be the Economic Impact
suggests, the online labour market demand of COVID-19 in the US? Rough Estimates of Disease
has recovered, and in some occupational areas Senarios. National Bureau of Economic Research
has surpassed, relative to pre-pandemic levels. Working Paper 26867.
Bedford, J., E. Delia, J. Giesecke, D. Heymann,
Acknowledgements C. Ihekweazu, G. Kobinger, H. Clifford Lane,
Z. Memish, M. Oh, A. Sall, K. Ungchusak & L.
We thank Raghav Raheja (Syracuse University) and Wieler (2020), COVID-19: towards controlling of
Gabrielle Vaccaro (Skidmore College) for their cen- a pandemic. The Lancet 395, pp. 1015–1018.
tral role in the qualitative data collection efforts. Beerepoot, N., & B. Lambregts (2015),
Likewise, we would like to acknowledge Dr. Otto Competition in Online Job Marketplaces: Towards
Kässi for engineering most of the OLI data collec- a Global Labour Market for Outsourcing Services?
tion infrastructure. Global Networks 15, pp. 236–255.
Braesemann, F., V. Lehdonvirta & O. Kässi (2020), Lehdonvirta, V. (2018), Flexibility in the Gig
ICTs and the Urban-rural Divide: Can Online Economy: Managing Time on Three Online
Labour Platforms Bridge the Gap? Information, Piecework Pplatforms. New Technology, Work and
Communication, & Society, 1–21. https://doi. Employment 33, pp. 13–29.
org/10.1080/1369118X.2020.1761857 Lehdonvirta, V., O. Kässi, I. Hjorth, H. Barnard
Bunyaratavej, K., J. Doh, E.D. Hahn, A.Y. Lewin & & M. Graham (2019), The Global Platform
S. Massini (2011), Conceptual Issues in Services Economy: A New Offshoring Institution Enabling
Offshoring Research: A Multidisciplinary Review. Emerging-economy Microproviders. Journal of
Group & Organisation Management 36, pp. 70–102. Management 45, pp. 567–599.
Clement, J. (2020), Interest in Remote Working McKay, C., E. Pollack & A. Fitzpayne (2019),
Related Search Terms during COVID-19 Outbreak Automation and a Changing Economy. The
March 2020. Statista. Available at <https://www. Aspen Institute Future of Work Initiative, April
statista.com/statistics/1105696/remote-working- 2019. Available at <https://assets.aspeninstitute.
google-searches-worldwide-during-covid-19-outbr org/conte n t/uploa d s/2019/04/Autom a tion
eak/>. Accessed on 1 June 2020. -and-a-Changing-Economy_The-Case-for-Action_
Corporaal, G.F. & V. Lehdonvirta (2017), Platform April-2019.pdf>. Accessed on 19 April 2020.
Sourcing: How Fortune 500 Firms are Adopting Online McKibbin, W. J., & Fernando, R. (2020) The
Freelancing Platforms. Oxford: Oxford Internet Global Macroeconomic Impacts of COVID-19:
Institute. Seven Scenarios. CAMA Working Paper 19/2020,
del Rio-Chanona, R. M., P. Mealy, A. Pichler, February 2020.
F. Lafond & D. Farmer (2020), Supply and Mokhtarian, P.L. (1991), Defining Telecommuting.
Demand Shocks in the COVID-19 Pandemic: An Transportation Research Record 1305, pp. 273–281.
Industry and Occupation Perspective. arXiv pre- Normile, D. (2020) Coronavirus cases have dropped
print arXiv:2004.06759. sharply in South Korea. What’s the secret to its
Dunn, M. (2017), Digital Work: New Opportunities success?. Science. http://dx.doi.org/10.1126/
or Lost Wages? American Journal of Management 17, science.abb7566
pp. 10–27. Ouchi, W. (1979), A Conceptual Framework for the
Galperin, H., & C. Greppi (2019), Geographical Design of Organisational Control Mechanisms.
Discrimination in Digital Labour Platforms. In: M. Management Science 25, pp. 833–848.
Graham, ed., Digital economies at global margins, pp. Pesole, A., M.C. Urzí Brancati, E. Fernández-
295–318. Cambridge, MA: MIT Press. Macías, F. Biagi & I. González Vázquez (2018),
Horton, J.J. (2010), Online Labour Markets. In: Platform Workers in Europe. JRC Science for Policy
A. Saberi, ed., Internet and Network Economics, pp. Report. Brussels: European Commission.
515–522. Berlin Heidelberg: Springer. Stephany, F., N. Stoehr, P. Darius, L. Neuhäuser,
Huws, U., W.B. Korte & S. Robinson (1990), O. Teutloff & F. Braesemann(2020), The
Telework: Towards the Elusive Office. Hoboken, NJ: CoRisk-Index: A Data-mining Approach to
John Wiley & Sons. Identify Industry-specific Risk Assessments
International Labour Organisation (ILO) Related to COVID-19 in Real-time. arXiv preprint
(2016), Non-standard Employment around the World: arXiv:2003.12432.
Understanding Challenges, Shaping Prospects. Geneva: Thunstrom, L., S. Newbold, D. Finnoff, M.
ILO Publications. Ashworth & J.F. Shogren (2020), The
International Labour Organisation (ILO) Benefits and Costs of Using Social Distancing
(2020), COVID-19 and the World of Work: Country to Flatten the Curve for COVID-19. Journal of
Policy Responses. Geneva: ILO Publications. Benefit-Cost Analysis. https://doi.org/10.2139/
Kalleberg, A.L. (2003), Flexible Firms and Labour ssrn.3561934.
Market Segmentation: Effects of Workplace Wenham, C., J. Smith & R. Morgan (2020), COVID-
Restructuring on Jobs and Workers. Work and 19: The Gendered Impacts of the Outbreak. The
Occupations 30, pp. 154–175. Lancet 395, pp. 846–848.
Kässi, O., & V. Lehdonvirta (2018), Online Labour Wood, A.J., M. Graham, V. Lehdonvirta & I.
Index: Measuring the Online Gig Economy for Hjorth (2019), Networked but Commodified:
Policy and Research. Technological Forecasting and The (Dis)Embeddedness of Digital Labour in the
Social Change 137, pp. 241–248. Gig Economy. Sociology 53, pp. 931–950.

Article 2 - DISTANCING BONUS OR DOWNSCALING LOSS

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Article 2 - DISTANCING BONUS OR DOWNSCALING LOSS

Uploaded by

Copyright:

Available Formats

DISTANCING BONUS OR DOWNSCALING LOSS?

THE CHANGING LIVELIHOOD OF US ONLINE

Received: April 2020; accepted May 2020

INTRODUCTION writing and editing, and other types of cognitive

An increase in the number of registered workers’ experiences, we draw on data from

Table 1. Summary statistics.

Job classification Gender

Administrative 32% Female 61%

Hours a week spent looking for work

Role of Freelance Work Min 0

Age Length on digital platforms (in years)

disparities found in freelance work. These im- Notes

You might also like