1

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

Work on Demand

I n June 2006, tech journalist Jeff Howe noticed that the Internet was
reshaping the boundaries of work.The long-standing distinction between
employees working for large companies and amateurs engaging in their
craft as a hobby had become increasingly blurred:
Hobbyists, part-timers, and dabblers suddenly have a market for their efforts,
as smart companies in industries as disparate as pharmaceuticals and television
discover ways to tap the latent talent of the crowd. The labor isn’t always free,
but it costs a lot less than paying traditional employees. It’s not outsourcing; it’s
crowdsourcing.1
Crowdsourcing, he subsequently explained, is the ‘act of taking a job trad-
itionally performed by employees and outsourcing it to an undefined, gen-
erally large group of people in the form of an open call’.2 Its rise from
nowhere has been little short of meteoric: a mere decade later, recourse to
the crowd has begun to permeate our daily lives. In cities around the world,
consumers can hail Ubers instead of traditional taxis, order their food
through Deliveroo, request handyman assistance from TaskRabbit, and out-
source small digital tasks on Amazon’s Mechanical Turk (MTurk). Welcome
to the gig economy.
The ramifications are far-reaching.Traditional companies are replaced by
platforms; their long-term employees recast as independent entrepreneurs.
The platform economy, Professor Orly Lobel of the University of San Diego
argues, is ‘not only a paradigmatic shift for business, but also for legal the-
ory’.3 Such bold claims raise a number of issues: are they true? And, in any
event, what is this idea of ‘on-demand work’ in the ‘gig economy’?
In this chapter, we set out to explore how the gig economy works: we
will meet some of the most important platforms, and illustrate their central
role in shaping transactions between consumer and workers. This digital
work intermediation is key to understanding the gig economy: platforms’

Humans as a Service: The Promise and Perils of Work in the Gig Economy. First Edition.
Jeremias Prassl. © Jeremias Prassl 2018. Published 2018 by Oxford University Press.
12 Wor k on De m a n d

sophisticated algorithms connect workers and customers, and exercise ongoing

control over the ensuing relationships.We then chart the astonishing variety
and global growth of the gig economy, with a particular emphasis on the
underlying business model: how do platforms make money? Finally, we turn
to the broader impacts of digital work intermediation: how do platforms go

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

beyond mere matchmaking to shape the experiences of workers and con-
sumers? The potential advantages of gig work are manifold, from the flexi-
bility to choose when and where to work, to the ability to offer innovative
and affordable services. But a life of intermittent, short-term ‘gigs’ also has its
downsides, from economic insecurity to a lack of traditional employment
protection.

Understanding the Gig Economy

Whether you are looking for food delivery from a nearby restaurant or a
lawyer to help with a complicated property dispute, it’s likely that there’ll
already be ‘an app for that’. An exhaustive list, let alone description, of gig-
economy platforms would far exceed the pages of this book—and be out-
dated before it could be printed. Despite their public prominence, platforms
such as Uber, MTurk, and TaskRabbit are but the tip of an iceberg. Each
business model has been copied and developed by a vast number of rival
start-ups: Uber faces competition in the United States from Lyft, Didi in
China, Ola in India, and a host of local taxi apps across the globe. Whereas
TaskRabbit tackles all sorts of odd jobs, a growing number of operators focus
on particular industries, such as restaurants: food-delivery rivals Deliveroo and
Foodora had expanded to more than ten countries by 2016.4 And MTurk faces
competition online, whether it’s Fiverr for quick jobs or platforms such as
Upwork for higher-skilled work, including digital design and programming.
A further complication arises from the fact that gig-economy brands
and business models vary across countries and operators. New platforms are
launched and old ones dispatched into bankruptcy on a daily basis. Even
well-established operators seem to reinvent themselves constantly: a plat-
form might experiment across different cities with its customer pricing, the
commission it charges, and how workers are assigned to particular tasks.
Indeed, even the descriptions in this chapter will have to resort to generaliza­
tions and are by no means future-proof: changes to the business model
require little more than a software update.
 U n de rsta n di ng t h e Gig Econom y13

To grapple with the constant evolution of the on-demand economy, sev-

eral academics have attempted to come up with a taxonomy of gig-economy
work. One of the leading authors in this field is Jan Marco Leimeister of
Kassel University in Germany. Together with his team, he has developed
an extensive classificatory scheme of crowdsourcing and crowdwork. This

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

distinguishes, for example, between ‘internal’ and ‘external’ crowdwork,
depending on whether on-demand workers are employed by the platform
operator or not.The latter category is then subdivided into a series of ‘arche-
types’, including categories such as ‘Microtask’, where ‘tasks are predomin-
antly simple and repetitive’, or ‘Marketplace platforms’, through which ‘more
long-term and complex jobs are given into crowd’.5
This understanding of crowdwork as a purely digital form of on-demand
labour, whereby tasks can be completed behind a computer anywhere, is
usually contrasted with gigwork, whereby tasks mediated through a platform
have to be completed offline—think food delivery or cleaning.6 The axes
along which the field could be organized are nearly unlimited. Whether
work is completed online or offline is one possible dimension, but it isn’t the
only differentiation by any means: think, for example, about the difference
between task-specific platforms (Lyft, Deliveroo) versus generalist operators
(Fiverr, TaskRabbit), or differences in who sets the price of each task (the
platform, in some cases; the consumer or even the worker herself, in others).7
In reality, all gig-economy platforms’ business models overlap and intersect—
with several crucial commonalities, including the use of algorithmic rating
mechanisms, and a notable fondness for coming up with new names and
labels.

Digital Work Intermediation

Even more importantly for present purposes, the gig-economy platforms’
business model is universally based on near-instant recourse to a large pool of
on-demand workers: the ‘crowd’, looking for their next ‘gig’.The gig econ-
omy’s promise to deliver speedy services at low cost is premised on the
availability of a large supply of workers who can easily be matched with
consumer demand. Behind the language of ‘gigs’, ‘tasks’, and ‘rides’ sits a much
more sophisticated business model: the digital intermediation of work.
At first glance, platforms are but a small cog in the gig economy, pairing
consumers looking for a particular service with a suitable entrepreneur
willing to complete the task. Algorithms take into account a wide range of
14 Wor k on De m a n d

relevant factors, from the quality of previous work and current availability
to geographic location, and optimize each match—before charging a small
fee for the service.
Upon closer inspection, however, the platform’s role goes far beyond
mere match-making: it offers digital work intermediation. To deliver tightly

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

curated products and services to customers, gig-economy operators actively
shape the entire transaction by means of close control over their workers.
We will explore the mechanisms behind this control in Chapter 3; suffice it
to say here that whilst elements of work in the on-demand economy might
not look like traditional ‘9-to-5’ jobs with a single employer, the reality of
work is often a far cry from the freedom and independence of genuine
entrepreneurship. As a result, the services offered to consumers are consider-
ably more than a simple one-off match, from quality monitoring to pay-
ment facilitation.
How does this work in practice? We will encounter a host of different
platforms over the course of the coming chapters: there is a near-limitless
variation of business models out there.To get a sense of digital labour inter-
mediation, however, let’s begin by exploring three archetypical operators to
see the myriad ways in which platforms’ algorithms shape the relationship
between customer and worker.
Uber is ‘Everyone’s Private Driver’ in cities across the globe. Its app offers
customers (‘riders’) multiple types of service (such as UberX at the econ-
omy end of the spectrum and UberLUX for premium cars), depending on
location.8 Other platforms, such as TaskRabbit, offer a much broader range
of services. Accessed through an app or website, the company advertises
help with jobs ranging from moving home and furniture assembly, to clean-
ing and small repair works, in nearly 40 US cities, as well as London.9 Digital
remote work, finally, is the third gig-economy archetype. MTurk was one of
the earliest operators in this field, connecting consumers and businesses
(‘Requesters’) with workers (‘Turkers’) across the globe. Despite the variety
of services on offer, each platform operates as a digital labour intermediary:
matching consumer demand with workers from its pool—and exercising
close control over the entire relationship.
Gig-economy services can be accessed with hitherto unimaginable ease
(and often at much cheaper prices) once consumers have registered their
personal and credit card details with a platform. Rather than heading out to
the streets in search of a taxi, you can simply order a ‘ride’ at the touch of a
button in any city in which Uber drivers are active, selecting a desired pick-up
 U n de rsta n di ng t h e Gig Econom y15

location and category of service.The app connects the rider with a close-by
driver, whose location can be monitored via the app. Information, including
the driver’s name, photo, and car details, is shared with the rider to ensure
easy identification. Once in the car (or beforehand, if the customer wants a
price estimate), the rider enters her destination; an algorithm automatically

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

works out the route.
On TaskRabbit, consumers choose a task category, and specify when and
where the work should get done. An algorithm matches each request with
a shortlist of ‘Taskers’ for the customer to engage, providing information
about their experience and hourly rates. Sometimes, a ‘Quick Assign’ feature
might also be available, allocating tasks to the first available worker. Details
about the flat to be cleaned or items to be delivered are then arranged through
the app.
On MTurk, jobs are broken down into so-called human intelligence tasks
(HITs): a wide range of low-complexity, low-skill tasks, from identifying
objects in photographs and extracting product details from receipts, to com-
pleting surveys and transcribing audio or video clips. Requesters upload
these tasks to the platform and set a price (rarely more than a couple of US
cents), time frame, and required worker qualifications.
The platforms, finally, also handle all payment transactions and provide
feedback through worker (and sometimes consumer) ratings. Upon com-
pletion of each task, MTurk Requesters can approve or reject the work;
payment is processed by the platform, with Amazon deducting a commission
of 20–40 per cent, with a minimum of half a cent per hit.
An Uber passenger can simply jump out of the car at her destination—
the price is determined by the app and payment processed through the
card on file.The platform retains a commission of 20–25 per cent (depend-
ing on the city) and pays the rest to the driver. After each ride, passengers
and drivers are invited to rate each other through a five-star system, the
cumulative results of which become their ‘ratings’ displayed before future
rides.
TaskRabbit similarly handles invoicing and payment on the basis of how
many hours a task took to complete. Each Tasker can set her own hourly
rates, to which the platform adds its commission (30 per cent, at the time of
writing) and a ‘Trust & Support Fee’ of 7.5 per cent to fund worker vetting,
customer support, and the platform’s US$1 million insurance guarantee should
‘something unexpected happen during the task’.10 After completion of the
task, clients and workers both leave feedback, with the client rating each
16 Wor k on De m a n d

worker out of five stars in response to a number of questions. A cumulative

rating, together with any public review comments, will be displayed on Taskers’
profiles to guide future customers.11

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

How Big Is the Gig Economy?
The platform economy, breathless futurologists assure us, is the future of
work: with ‘freelancing [as] the new normal’,12 it will fundamentally reshape
the organization of businesses, the economy, and our working lives. Not
everyone agrees. Frank Kalman, editor of Talent Economy magazine, is ‘not
buying it’.13 The gig economy, he argues, represents a tiny fraction of our
labour markets, goes against the grain of corporate work culture, and imposes
a host of hidden coordination and transaction costs on traditional businesses.
In short, ‘gig work is likely to remain a small part of the overall labor force,
both from an economic perspective and a cultural, performance and man-
agement perspective’.14
How big, then, is the gig economy? Depending on where we look, we
are faced with wildly different numbers—especially when trying to deter-
mine what proportion of the overall workforce are engaged in the gig
economy.15 Very little, some argue: economists Lawrence Katz of Harvard
University and Alan Krueger of Princeton University, for example, esti-
mated in 2016 that a mere 0.5 per cent of the US workforce worked for
on-demand platforms—that is, no more than 800,000 workers.16 US
Senator Mark Warner, meanwhile, cites a much larger (if hardly credible)
range of estimates: ‘I've seen it range from 3 million to 53 million.’17
The truth lies somewhere in between those extremes. Indeed, a more
realistic consensus (at least as regards industry size) appears to be emerging
at the time of writing. Several studies using a range of methodologies, from
traditional surveys to an analysis of bank accounts to determine where
income is derived from, have homed in on a figure of approximately 4 per
cent of the working-age population both in the United States and the
UK.18 A report by the RSA, a UK think tank, published in spring 2017 simi-
larly estimates that there are currently 1.1 million gig workers in the UK
and that approximately ‘3 per cent of adults aged 15+ have tried gig work
of some form, which equates to as many as 1.6 million adults’.19
From an overall labour-market perspective, these numbers don’t necessarily
sound like a major concern—until we consider the fact that most serious
attempts at measuring the size of gig work in the broader labour market
 U n de rsta n di ng t h e Gig Econom y17

tend to understate its extent. Current statistical measures often fail to take
into account the full scope of gig-economy work, not least because they
tend to focus on primary income sources; workers supplementing their
income with gig-economy work are thus likely to be excluded from official
statistics.20

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

The gig economy is also an increasingly global phenomenon. Digital
work, in particular, can easily be outsourced across borders. In January 2017,
colleagues at the Oxford Internet Institute published the results of an inten-
sive three-year study into online gig work across the globe. Their conclu-
sion is clear: platform work ‘is becoming increasingly important to workers
living in low- and middle-income countries’.21 Their detailed analysis of gig
work in Kenya, Nigeria, South Africa,Vietnam, Malaysia, and the Philippines
paints an already-familiar picture:
[R]egions like Sub-Saharan Africa and Southeast Asia, in particular, can capit-
alise on this digitally-mediated work opportunity. New sources of work are
especially needed as the youth-to-adult unemployment rate hits historic peaks
and average wages remain significantly lower in emerging economies than in
developed economies . . . [But there are] also concerns such as downward pres-
sure on pay, long hours, discrimination, and lack of social contact for some.22

From London to Cape Town, New York to Hanoi, the opportunities and
threats we observe mirror the tensions we will encounter throughout this
book.

Explosive Growth
Despite our problems in pinning down the exact size of the gig economy
today, given its clear—and global—growth trend (the figures cited above will,
in all likelihood, be long outdated by the time you read them), on-demand
work is likely to become an ever-more-salient topic in years to come.23 The
industry is booming. As Micha Kaufman, CEO of task platform Fiverr,
noted as early as 2013:
[A] revolution is taking shape—an entirely different kind of economy. The
labor force of new entrepreneurs, which we call the Gig Economy, is growing
rapidly around the world and could soon represent as much as 50 per cent of
the US workforce.24

Whether it’s the range of tasks offered and industries affected, turn­over,
or the numbers of consumers and workers, whichever way you measure
the gig economy, growth rates are enormous.25 Growing consumer demand
18 Wor k on De m a n d

is a key factor driving industry growth: as more and more tasks become
available through online apps, both individual consumers and business users
will come to rely on on-demand work.26 Home-cleaning platform Handy,
for example, grew to bookings worth US$1 million per week within two
years of its launch.27 Uber’s app has proved similarly popular with consumers.

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

Founded in San Francisco in the spring of 2009, less than a decade later
Uber was operating in nearly 600 cities across 81 countries—providing
more than 15 million journeys around the world on New Year’s Eve 2016
alone.28
This explosive growth is fuelled by eye-watering amounts of venture capital
investment: whereas a mere US$57 million was invested in on-demand
platform start-ups in 2010, four years later that sum had risen to more than
$4 billion. By 2017, investors had poured more than $12 billion into Uber
alone—giving the company a valuation of nearly $70 billion.29
Why is the gig economy so attractive to investors? To answer this ques-
tion, we need to return to the platforms’ business model and take a deeper
look at the underlying economics. In a world of digital work intermediation,
how do the intermediaries actually make their money? And why are many
operators valued so highly by venture capital firms and other sophisticated
investors?

The Economics of the Gig Economy

Delving into the figures, we quickly discover a number of possible explan­
ations. Some argue that the gig economy creates value through the platforms’
faster matching of consumer demand and worker supply, relying on clever
algorithms and sophisticated rating systems to grasp otherwise wasted busi-
ness opportunities. Another school of thought is more critical, suggesting
that regulatory arbitrage and negative externalities are at the core of most
platforms’ valuations.

Matching and Intermediation

We have already seen how the dominant story behind the gig economy’s
commercial success is one of platforms’ vastly superior matching opportunities,
unlocking and creating surplus value in the economy. Platforms then reap a
 T h e Econom ics of t h e Gig Econom y19

small percentage of this added value through their fees.There is undoubtedly

a core of truth to this: faster matches greatly improve the functioning of
product and service markets.
Economics Nobel Laureate Christopher Pissarides was amongst the first
to formalize this model in the employment context.30 Imagine a labour

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

market in which some firms are looking for workers in one city and indi-
viduals are looking for jobs in another. It will be hard for either party to find
out about the other, leaving jobs unfilled and workers unemployed. These
‘search frictions’ are wasteful and leave everyone worse off. Through
technological innovation, from location tracking and user ratings, to
sophisticated algorithms that match consumers and workers, whether purely
online or in the real world, platforms have drastically reduced this friction:
using Amazon’s MTurk, an economist in Australia can quickly find a student
in the United States to help her to organize data in a large spreadsheet.
Even better, matching isn’t the only market friction removed by the
platform. In an open-market transaction with an independent entrepreneur,
consumers would have to spend significant amounts of time and effort to
find out information about the service provider’s background and experi-
ence, control the quality of the work, and negotiate prices. This is the real
value of digital work intermediation: gig-economy operators also provide
information about how reliable a worker is, take care of invoicing and pay-
ments, and provide a (digital) infrastructure within which the entire exchange
can take place.
With transaction cost so drastically reduced, the narrative continues, the
traditional firm as described by Ronald Coase becomes obsolete; instead,
we move into a hybrid world between markets and hierarchies. According
to Coase’s theory of the firm, companies exist because the control exercised
by an entrepreneur-coordinator over her workforce and other factors of pro-
duction is much cheaper than the cost involved in going out to the market
and haggling over each individual transaction.31 On the other hand, once an
app has taken all of the hassle out of such transactions, Coase’s entrepreneur
will no longer need to strike long-term bargains with workers, let alone
invest in assets; she can replace her workforce with an external crowd, ready
to complete individual tasks as and when required.
Better matching and lower transaction cost undoubtedly unlock value:
the economist finishes her job more quickly, the student earns some extra
income—and Amazon charges a few US cents’ commission for its matching
and intermediation service. So far, so good: the Internet is used to facilitate
20 Wor k on De m a n d

transactions between strangers that could otherwise not have taken place.
That story, however, fails to answer a fundamental question: why are shar-
ing-economy platforms worth so much? How can start-up businesses that
simply rely on matching supply and demand via the Internet quickly be
valued in the millions or billions and lay claim to an economic revolution?

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

Digital matching services, online marketplaces, and classified sites such as
Craigslist and Gumtree have, after all, been around for nearly two decades—
and are valued at much lower prices.

Regulatory Arbitrage
Professor Julia Tomassetti is highly critical of the suggestion that platforms’
primary value creation is achieved through better matching and lower
transaction cost: ‘What happens when we actually subject the Uber narra-
tive to scrutiny under Coasian theory? It does not hold up. From the
Coasian perspective, Uber does not write the epitaph of the firm.’32 Platforms,
she argues, speak the language of markets—but they operate like old-­
fashioned employers, relying on technology to exercise tight control over
their workforce.
Tomassetti doesn’t deny that gig-economy platforms have dramatically
lowered transaction cost in comparison with established competitors. Lowering
transaction cost alone, however, cannot account for platforms’ phenomenal
valuations and claims to disruptive innovation: there is, despite all claims to
the contrary, little that is genuinely novel as far as platforms’ production pro-
cesses are concerned. Uber follows the basic lines of a traditional taxi firm;
TaskRabbit, those of a labour-outsourcing agency. The key to understanding
the business model, Tomassetti points out, is a different one: platforms are
but the latest example of ‘postindustrial corporations’.33 They seek ‘to maxi-
mize profit, but not necessarily through productive enterprise. Rather, [they]
may create shareholder value by other means, like asset manipulation, specu-
lative activity, and, most pertinent here, regulatory arbitrage.’34
What does that mean? Victor Fleischer’s seminal work defines regulatory
arbitrage as ‘the manipulation of the structure of a deal to take advantage of
a gap between the economic substance of a transaction and its regulatory
treatment’.35 Firms, in other words, may try to structure their business so as
to hide what is actually going on from regulators and evade the law. His first
example of a ‘pervasive’ arbitrage technique? ‘[F]iring employees and re-
hiring them as independent contractors to avoid employment regulation.’36
 T h e Econom ics of t h e Gig Econom y21

In that sense, then, employment law—or rather the evasion of employ-

ment law—is at the core of the gig-economy business model. Recourse to
large pools of on-demand workers is ‘the economic substance’ of platforms’
transactions. Employing a large workforce creates responsibilities—which
impose cost on people-intensive business models. In return for the benefits

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

of control over their workforce, employment regulation places a financial
burden on employers, ranging from social security contributions, minimum
wage laws, and sick leave, to health and safety regulations and union bar-
gaining. Stable employment relationships are also associated with indirect
cost, because the risk of fluctuations in demand cannot be offloaded onto
individual workers: a bus company’s drivers ply their routes and receive
wages regardless of whether passengers are on board or not.37
Regulatory arbitrage in the gig economy takes many forms: think about
ride-sharing platforms’ insistence that taxi regulation does not apply to their
business, for example. Portraying workers as independent entrepreneurs and
refuting their employment status, on the other hand, is a consistent theme
throughout: shareholder value is created by denying workers their legally
mandated rights. Classifying workers as independent contractors allows plat-
forms to offer services without having to pay for their cost. Responsibility
for assets, remuneration, insurance, and tax, as well as the risks of fluctuating
demand, are devolved to individual micro-entrepreneurs. The potentially
enormous gains from this financial arbitrage are at the core of gig-economy
business models. In the words of one financial adviser, ‘adverse determin­
ations in these matters may subject [the platform] to additional compensation
expenses or taxes in certain jurisdictions, which could have a material adverse
effect on its ability to operate its business’.38
Regulatory arbitrage also leads to negative externalities: the social cost of
platforms’ activities are higher than their private cost.Think, for example, of
a number of ride-sharing cars roaming the streets whilst looking for the
next passenger. We have already seen that platforms usually try to have as
many workers as possible available at any given time.This is enticing to con-
sumers, who can quickly catch a ride in whichever car is closest—whilst the
drivers suffer precisely because supply is designed far to outstrip demand.The
platforms’ algorithms have trapped workers in a low-productivity environ-
ment in order to fulfil orders as quickly as possible, whilst ignoring the costs
incurred by the other drivers looking for a job—from the cost of their
time and petrol, to the environmental implications of having a large number
of cars polluting away. Classic economic theory spells out the implications:
22 Wor k on De m a n d

if a service provider doesn’t have to bear the full range of costs created by her
product or service, she will end up offering too much of it. The platform
always wins, even if individual workers—and indeed society at large—lose out.

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

Cash Burn
On the basis of what we have seen so far, gig-economy shareholders should
be able to rejoice and reap considerable profits: growing demand can be met
without capital investment or other downside risks. In reality, however, most
operators have been sustaining high losses. According to transportation indus-
try analyst Hubert Horan: ‘[I]n the year ending September 2015, Uber had
GAAP [Generally Accepted Accounting Principles] losses of [US]$2 billion
on revenue $1.4 billion, a negative 143% profit margin ... 2016 GAAP losses
would easily exceed $3 billion.’39 Numbers are less extreme for other plat-
forms, but the underlying struggle for profitability is the same. TaskRabbit,
for example, had long aimed to turn a profit by the end of 2016—and promptly
‘backed away from that claim’ when the time came, whilst continuing to
insist that it would ‘get there soon’.40
Horan suggests that subsidies funded by venture capital are a major factor
behind these losses:
Uber passengers were paying only 41% of the actual cost of their trips; Uber
was using these massive subsidies to undercut the fares and provide more cap-
acity than the competitors who had to cover 100% of their costs out of pas-
senger fares. Many other tech startups lost money as they pursued growth and
market share, but losses of this magnitude are unprecedented.41

As gig-economy platforms become mature market players, mounting losses

raise a difficult question: will they ever be profitable? Or has regulatory
arbitrage camouflaged a different model altogether, leading to a misalloca-
tion of capital? Writing for the Financial Times, Izabella Kaminska was
amongst the first to raise these concerns as part of her long-running inves-
tigation of Uber’s financials:
If Uber is cheap it is not because it has out innovated the incumbent cab
market, which at the end of the day has access to exactly the same ride-
hailing technology. To the contrary, it’s because investors have failed to rec-
ognise that the source of its greatest innovations is and always has been cheap
money.
Indeed, from egregious undercutting tactics based on promotional give-
aways to turning a blind eye to exploitative labour practices thanks to the
 T h e Econom ics of t h e Gig Econom y23

cheap funding of aggressive lobbying campaigns aimed at changing legal

frameworks or the reckless flooding of the market with huge amounts of spare
capacity, none of it would be possible without access to cheap financing.42

If Kaminska is correct, we are still left with one final question: assuming that

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

the rapid growth of gig-economy platforms has really been fuelled by little
more than a combination of regulatory arbitrage and cheap venture capital,
why are savvy investors competing to invest? What is behind their willing-
ness to burn unprecedented amounts of cash?

Network Effects and Monopoly Power

It is common for start-up businesses to lose money early on, of course—not
least by subsidizing products so as to gain market share. This is particularly
important in industries in which investors hope to harness so-called net-
work effects—that is, where all users of a particular service gain if additional
consumers adopt it.43 Think about the growth of a ride-sharing platform in
a new city, for example. If a large number of consumers are using a particu-
lar app to hail taxis, it will become more attractive for drivers to sign up to
that app. A large available pool of drivers, in turn, will make it easier and
cheaper for consumers to find their next ride, further increasing the incen-
tives for new drivers to join—and so on. It is unsurprising that gig-economy
platforms will often try to kick-start this process by investing significant
amounts of cash in subsidies for drivers as well as passengers.
Hubert Horan, however, is sceptical that this is the entire story. Cash burn,
he suggests, is not merely about harnessing network effects, but rather a step
in platforms’ quest for monopoly power. Focusing once more on Uber as
the most pointed example, he explains the links:
[M]ost critically, the staggering $13 billion in cash its investors provided is
consistent with the magnitude of funding required to subsidize the many years
of predatory competition required to drive out more efficient incumbents.
Uber’s investors did not put $13 billion into the company because they
thought they could vanquish those incumbents under ‘level playing field’ mar-
ket conditions; those billions were designed to replace ‘level playing field’
competition with a hopeless battle between small scale incumbents with no
access to capital struggling to cover their bare bone costs and a behemoth
company funded by Silicon Valley billionaires willing to subsidize years of
multi-billion dollar losses. Given Uber’s growth to date, investor expectations
that monopoly rents justifies the current level of subsidies and financial risks
appears quite plausible.44
24 Wor k on De m a n d

This account stands in stark contrast with the idea that the rise of gig-
economy platforms will lead to increased competition, with lower prices
and higher quality as the result: in the expensive pursuit of network effects,
some platforms’ goal may well be to smother competition, rather than to
encourage it.

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

Individual operators will always vary in the extent to which their busi-
ness model combines the factors thus set out. Whichever way we approach
the question, however, the underlying economics don’t appear to stack up—a
quandary to which we return in the final chapter. Faster matching, digital
work intermediation, and assorted rating algorithms have the potential to
create much economic benefit. Regulatory arbitrage, externalities, and lack-
ing profitability despite considerable cash burn, on the other hand, should
give us all pause for thought.

The Promise—and the Perils—of

On-Demand Work
For the time being, however, most of us appear to pay scant notice to these
economic controversies. Businesses and consumers alike are turning to the
gig economy to find rides across town, get errands and odd jobs sorted,
and crowdsource work online, whilst a seemingly ever-increasing pool of
workers competes to meet the demand. As a result, both consumers’ and
workers’ day-to-day experiences of life and work in the gig economy have
increasingly been subjected to public scrutiny—with decidedly mixed results.

Looking on the Bright Side of Life

Senior industry executives argue that, in times of financial crises and rising
unemployment, gig-economy platforms ‘are glimmers of hope . . . the force
that saves the American worker’.45 Viewed thus, the gig economy offers
incredible opportunities to workers: it provides flexible work, and a chance
to earn additional income as and when needed,46 without the regimented
working day and overbearing boss typical of traditional work. It can be a
real lifeline for groups traditionally excluded from the labour market, from
homebound workers to those with criminal convictions. Anyone can easily
find work through digital intermediary platforms such as Fiverr or Amazon’s
 T h e Prom ise —a n d t h e Pe r i ls— of On-de m a n d Wor k25

MTurk, where an online account is all that’s required to start earning

money.
Many workers agree—and are happy to share why they enjoy working in
the gig economy. Uber’s website, for example, features a series of ‘Driver-
Partner Stories’:

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

Mothers and fathers. Teachers and students. Artists and athletes. Uber driver-
partners come from all walks of life.Whether they drive with Uber to support
their passion, or just to earn extra cash in their spare time, meet Uber partners
from around the world and learn what moves them.47

Their stories are genuinely inspiring: the flexibility and additional income
of gig work allow many to pursue their goals and fulfil long-held dreams.
Meet Yoseph, who emigrated from Ethiopia to New Haven, Connecticut,
and fits in Uber driving around his university classes: ‘What Yoseph loves
most about driving in New Haven is meeting riders. “Yale attracts all kinds
of interesting people”.’48 Or ‘rockstar mom and Uber Partner, Christine,
who drives to balance marketing for the family business and being in the
stands for all of her son’s baseball games’. Her advice? ‘Do it!’49 And then
there’s Loren: artist, Seattle school bus driver, adventurer. ‘Uber has allowed
her to work flexible hours whenever she wants to “and if you need to take
a nap or get groceries you just turn off the blue button and come back to it
whenever you want”.’50
It is not only workers who are better off. Consumers, too, profit from the
rise of the gig economy. Previously unaffordable services are suddenly avail-
able at the touch of a button. Anyone who has hailed an Uber or commis-
sioned a survey on MTurk will be familiar with the ease and satisfaction that
comes from speedy and low-cost service delivery. Studies of consumer pref-
erences for ride-sharing platforms reveal the joys of cheaper pricing, wider
choice, and easier access.51
There’s nothing that seems impossible with the help of gig-economy
platforms—even if that means organizing a kid’s birthday party ‘on one of
the top rated beaches in all of SoCal!’:
With about 40 expected guests for her son’s 8th birthday beach party, a Client
in Los Angeles was, understandably, looking for a little assistance ... Luckily,
TaskRabbit was able to connect her with Tasker Rain F. Rain has experience
working with kids, and she’s also an accomplished event planner and staffer—
the perfect combo for this task. She did it all with a smile and, with her help,
the party was a big success!52
26 Wor k on De m a n d

The benefits extend to society at large: the gig economy allows us to tap
into otherwise wasted time and assets, driving economic growth. Platforms
create jobs in a world of rising unemployment, provide extra revenues to
workers, and formalize work in traditionally problematic sectors. Their
matching algorithms can foster ‘greater trust between strangers ... a corner-

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

stone of transactions, technology being the other’.53
The rise of gig work might also have more indirect and unexpected benefits
to society: a study by Jessica Lynn Peck of the City University of New York
argues that the introduction of Uber has a clear effect on incidents of drink-
driving: since the platform’s arrival in New York City in May 2011, she
estimates, there has been a 25–35 per cent decrease in alcohol-related
collisions.54 On-demand work might even benefit the environment: a 2010
study of car-sharing in North America suggested that for each shared
vehicle, 9–13 cars could be taken off the road.55

The Dark Side of the Moon

Just as there is much to be lauded and enjoyed about the on-demand econ-
omy, however, there is also cause for concern. A growing number of articles,
policy reports, and academic writings present another—darker—picture. The
gig economy, this line of thinking suggests, cynically exploits workers—and
might also be harmful to consumers and the public at large.
Until the UK’s general election in the early summer of 2017, Frank Field
MP chaired the House of Commons Work and Pension Committee, over-
seeing a major enquiry into gig-economy work. The evidence before the
Committee was troubling. In a private report subsequently submitted to
Transport for London, he urged the regulator to address ‘three evils’ and
‘insert a much needed floor into the bottom of this particular area of the
labour market’:
We are troubled to note that the practices described to us by drivers working
with Uber would appear to fit the Victorians’ definition of ‘sweated labour’. In
1890, a House of Lords Select Committee deemed labour to be ‘sweated’
when earnings were barely sufficient to sustain existence; hours of labour were
such as to make the lives of the workers periods of almost ceaseless toil; and
conditions were injurious to the health of the workers and dangerous to the
public.56

An analysis of drivers’ submissions suggested that drivers were ‘at risk of tak-
ing home less than a third of the National Living Wage’, even though they
 T h e Prom ise —a n d t h e Pe r i ls— of On-de m a n d Wor k27

had to ‘stay on the road for extended periods of time to make a living’ and
did ‘not have the freedom to determine their own working patterns’.57
The US National Employment Law Project (NELP) has been a similarly
vocal critic of gig-economy work, highlighting ‘micro wages’ and exploit­
ative working conditions.58 Unpredictable consumer demand makes flexi­

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

bility illusory; invasive data collection denies users privacy; workers are
increasingly isolated and left to compete against each other. The NELP’s
report into the on-demand economy cites an online posting by a frustrated
crowdworker:
Horrific. Digital sweat shop, slave wages, sometimes NO wages. You will be
asked to jump through an absurd number of hoops for less than minimum
wage. If you have a college degree and are either a professional writer, or a
professional in the field you are writing about, don’t even lower yourself to
this. It only kills your professional self worth.59

A survey by the UK’s Chartered Institute of Personnel and Development

(CIPD) further warned of the overall lower financial resilience of gig-economy
workers, nearly half of whom said that they could only ‘afford to live for
either less than one month or for up to two months without falling behind
on paying key bills or living expenses if they lost their income due to unfore-
seen circumstances’.60 Psychologists increasingly highlight the dangerous
public health implications of low-paid, insecure work.61
These concerns are nothing new: scholars have long argued that technol-
ogy is a major driver of labour-market inequality, giving rise to a ‘cyber-
proletariat’.62 Labour sociologist Professor Ursula Huws wrote about ‘the
making of a cybertariat’ several decades before the advent of the gig econ-
omy.63 Rapid technological advances, however, mean that many of these
trends have found their sharpest edge yet—with worries about consumer
and broader societal welfare similarly on the rise.
Consider the promise of better service at cheaper prices, for example—a
picture increasingly contradicted by quality concerns about gig-economy
products and services. It’s not difficult to see why quality might suffer: ‘[T]o
be hyper-affordable these are not necessarily on a par with established and
leading competitors.’64 Prices, too, are often more expensive than consumers
are led to believe: dynamic algorithms allow platforms to increase prices in
line with demand and a wide range of other factors—up to many multiples of
the underlying cost.65
As regards society at large, it is hard to deny the overall gains of increased
economic activity—but how equally are these distributed? Commentators
28 Wor k on De m a n d

are concerned about the entrenchment of a two-tiered labour market. As

The Economist has noted, gig-economy entrepreneurs:
have created a plethora of on-demand companies that put time-starved urban
professionals in timely contact with job-starved workers, creating a sometimes
distasteful caricature of technology-driven social disparity in the process; an

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

article about the on-demand economy by Kevin Roose in New York magazine
began with the revelation that the housecleaner he hired through Homejoy
lived in a homeless shelter.66
Other concerns extend to the future of a competitive market economy at
large. Consider, for example, the RSA’s worry that:
[A] small number of sharing platforms have been able to scale their networks
to an extent where they are beginning to show signs of monopoly power in
influencing the price, output, and investment of an industry, as well as in limit-
ing the entry of new competitors.67

Some gig-economy platforms are not simply nimble start-ups; they could pose
a real threat of cornering markets and becoming entrenched monopolies.

What Is Going On?

Turning to the empirical evidence behind these competing accounts, we
reach a counter-intuitive conclusion: there is truth in both. How can that
be? Because of the vast heterogeneity underpinning gig-economy work, in
tasks, as well as working conditions. This diversity is one of the key themes
emerging throughout the book.68 We cannot think of gig work as a single,
homogeneous category, tempting though it might be to do so, particularly
when looking for ‘the solution’ to concerns about on-demand work.
Perhaps unsurprisingly, a recent World Bank report found that the het-
erogeneity of work translates into starkly different worker experiences:
A part-time Filipino online worker reported earnings of [US]$3–4 per hour
on oDesk (rebranded as Upwork), performing tasks such as transcription, data
entry, and basic administrative services. In contrast, an experienced Nigerian
online freelancer reported earnings of $20 per hour for software development
and website design. At the high end, online freelancers consulting on patents
or venture capital can earn more than $40 per hour.69

A detailed study by consulting group McKinsey in the autumn of 2016 div-

ided gig-economy workers into four categories: free agents, who ‘derive their
 W h at is Goi ng on ?29

primary income from independent work and actively choose this working
style’; casual earners, who choose to rely on gig work to supplement their
income; reluctants, who ‘derive their primary income from independent
work but would prefer traditional jobs’; and the financially strapped, who
‘would prefer not to have to do side jobs to make ends meet’.70 Overall, the

Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024

report suggests, a majority of the gig workforce falls into the first two cat-
egories, engaging in gig work as a preferred choice. The report’s authors
nonetheless observe that:
the magnitude of the problem is still striking. Scaling up the results of our
survey suggests that 50 million Americans and Europeans are independent out
of necessity, and more than 20 million of them rely on independent work as
their primary source of income.71

Even more importantly, the study found that low-income households were
much more likely to depend on insecure and unpredictable independent
work—as opposed to higher-income households, for whom it was a matter
of choice.72 Freedom of choice, it turns out, is also a key factor in determin-
ing whether a gig worker will enjoy her job: ‘Not surprisingly, those who
work ...because they feel their circumstances have forced them into it report
much lower levels of satisfaction.’73
Quantitative empirical evidence on the broader impacts of the on-
demand economy ‘is very partial and inconclusive—in many cases, it is
simply anecdotal and often presented by stakeholders in the current contro-
versies’.74 As a result, it is crucial that in discussing the gig economy we
keep both its promise and perils in mind. Subsequent chapters will delve
deep into the industry to explore why workers’ and consumers’ experiences
can vary greatly; for now, it is time to explore the broader implications of
the competing stories we have encountered—for they are crucial building
blocks in shaping narratives about, and thus the legal regulation of, the gig
economy.
Downloaded from https://academic.oup.com/book/26367/chapter/194689976 by Christopher Rääf user on 31 August 2024