Platform Economics: Rhetoric and Reality in the "Sharing Economy" 1787438104, 9781787438101

Platform Economics tackles head on the rhetoric surrounding the so-called "sharing economy", which has muddied

219 92 4MB

English Pages 216 Year 2018

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Contents
Lists of Figures, Tables and Boxes
Figures
Tables
Boxes
Introduction
A Rhetoric-Driven ‘Negative Policy Bubble’
Elective Affinity
Interpretative Framework: Sharing Rhetorics and Evidence-Based Policy
Sources and Structure of the Book
Chapter 1: Platform Economics and the Sharing Economy: A Primer
Introduction
In Search of Two- and Multi-sidedness
Classifications, Determinants of Size and Market Functioning
Size Determinants
Inherently Frictional Markets
Centralisation Versus Decentralisation
On-demand Versus Scheduled Transactions
Ratings and Related Governance Issues
Back to ‘Sharing Platforms’
Chapter 2: Rhetoric, Reality, Impacts and Regulation in Labour Intermediation Platforms
Introduction
Trajectory and Conceptual Issues
How Big is the Sharing Economy?
Rhetorical Discourses
Lobbying as Framing: Harnessing Rhetoric and Evidence
Box 1: Airbnb Self-reported Impacts.
Box 2: Uber’s Self-reported Impacts.
Social Capital and Motivation
Distributional and Stratification Effects
Environmental and Socio-economic Impacts
Environmental Impacts
Socio-economic Impacts
Consumer Welfare and Distributional Effects
Ratings and Platforms Functioning: Self-regulation?
From Legal Battles towards Regulation and Policy?
Box 3: Conflicts, Bans and Court Cases.
Chapter 3: Digital Labour Markets in a Broader Perspective
Introduction
Conceptualisation and Dimensional Relevance
Box 4: Online Micro-tasking Ideal-typical Functioning.
Box 5: Online Tasking, Ideal-typical Functioning.
Rhetorics and Economic Hypotheses
Descriptive Socio-economic Picture
Socio-demographic Contours of Labour Providers
Motivations
Employment Status
Working Conditions: Earnings and Other Aspects
New Evidence on Labour Providers in Context of NSW and Access to Social Protection
Evidence on Broadly Defined Market Efficiency Hypotheses
Matching Frictions, Market Inefficiencies and Biases
Super Star or Long Tail Effects?
Why Do Firms Hire from OLMs?
Net Aggregate Effects
Distributional Effects in MLMs
Digital Labour Markets in a Broader Perspective
Technological Trends: Digital Labour Markets and the Future of Work
Not Simply Technology
Non-standard Work and its Effects
From Legal Battles towards Regulation and Policy?
US Legal Disputes
Legal Disputes in Europe
The Debate
EU-level Context and Developments. There have been various regulatory and policy developments in many of the EU28 countries that would be beyond the scope of this final paragraph to review. We conclude instead by drawing the contours on EU-level general f
EU-level Flexibilisation of Work. NSW arrangements have been introduced and regulated widely at EU level with the three key directives for part-time (1997), fixed-term (1998) and temporary agency contracts (2008). As noted (Peers, 2013), the spirit of the
The EU Agenda on Collaborative Economy. In June 2016, the European Commission presented a European agenda for collaborative economy, a non-binding act of soft law that aims at adapting and interpreting the existing regulation, reassuring rights and obliga
European Parliament Resolutions. The European Parliament has been pushing to bring providers in digital labour markets more firmly into national social security systems amid concern that a changing labour market might leave them out from entitlements and
Chapter 4: Ideological Production in Digital Intermediation Platforms
Introduction
Integrated Theoretical Framework for In-depth Qualitative Ideological Production Analysis
Methodology: In-depth Interviews, Participant Observation, Secondary Document Analysis
‘Sometimes It’s Too Ideological’: The Challenge of Collaborative Players to Steer the ‘Conversation’
Against, with and Beyond State and Capital: Commons Discourses, Multifarious and Paradoxical
And in the Name of Platform Co-operativism: ‘We’re Interested in Exploring the Whole Spectrum of Options’
Conclusion
Chapter 5: Conclusions and Research Agenda
Introduction
From Rhetoric to Conceptual Clarity and Evidence
Evidence General
Policy and Regulation: Debate and Open Issues
Labour-specific Issues
Our View on Labour Issues
Lobbying as Rhetorical Framing: The Role of Evidence and a Research Agenda for the Future
References
Index
Recommend Papers

Platform Economics: Rhetoric and Reality in the "Sharing Economy"
 1787438104, 9781787438101

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

PLATFORM ECONOMICS

DIGITAL ACTIVISM AND SOCIETY: POLITICS, ECONOMY AND CULTURE IN NETWORK COMMUNICATION The Digital Activism and Society: Politics, Economy and Culture in Network Communication series focuses on the political use of digital everyday networked media by corporations, governments, international organisations (Digital Politics) as well as civil society actors, NGOs, activists, social movements and dissidents (Digital Activism) attempting to recruit, organise and fund their operations through information communication technologies. The series publishes books on theories and empirical case studies of digital politics and activism in the specific context of communication networks. Topics covered by the series include, but are not limited to: ⦁⦁ the different theoretical and analytical approaches of political communication

in digital networks;

⦁⦁ studies of socio-political media movements and activism (and ‘hacktivism’); ⦁⦁ transformations of older topics such as inequality, gender, class, power, iden-

tity and group belonging;

⦁⦁ strengths and vulnerabilities of social networks.

Series Editor Dr Athina Karatzogianni

About the Series Editor Dr Athina Karatzogianni is an Associate Professor at the University of Leicester, UK. Her research focuses on the intersections between digital media theory and political economy to study the use of digital technologies by new socio-political formations.

Published Books in this Series Digital Materialism: Origins, Philosophies, Prospects by Baruch Gottlieb Nirbhaya, New Media and Digital Gender Activism by Adrija Dey Internet Oligopoly: The Corporate Takeover of Our Digital World by Nikos Smyrnaios

Forthcoming Titles Digital Activism and Society (6)

PLATFORM ECONOMICS: RHETORIC AND REALITY IN THE ‘SHARING ECONOMY’ CRISTIANO CODAGNONE ATHINA KARATZOGIANNI JACOB MATTHEWS

United Kingdom – North America – Japan – India – Malaysia – China

Emerald Publishing Limited Howard House, Wagon Lane, Bingley BD16 1WA, UK First edition 2019 Copyright © 2019 Emerald Publishing Limited Reprints and permissions service Contact: [email protected] No part of this book may be reproduced, stored in a retrieval system, transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without either the prior written permission of the publisher or a licence permitting restricted copying issued in the UK by The Copyright Licensing Agency and in the USA by The Copyright Clearance Center. Any opinions expressed in the chapters are those of the authors. Whilst Emerald makes every effort to ensure the quality and accuracy of its content, Emerald makes no representation implied or otherwise, as to the chapters’ suitability and application and disclaims any warranties, express or implied, to their use. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: 978-1-78743-810-1 (Print) ISBN: 978-1-78743-809-5 (Online) ISBN: 978-1-78743-985-6 (Epub)

To our children and all future workers

This page intentionally left blank

Contents

Lists of Figures, Tables and Boxes Introduction

ix 1

Chapter 1 Platform Economics and the Sharing Economy: A Primer

17

Chapter 2 Rhetoric, Reality, Impacts and Regulation in Labour Intermediation Platforms

35

Chapter 3  Digital Labour Markets in a Broader Perspective

73

Chapter 4 Ideological Production in Digital Intermediation Platforms

123

Chapter 5  Conclusions and Research Agenda for the Future

151

References

169

Index

201

This page intentionally left blank

Lists of Figures, Tables and Boxes

Figures Fig. 1. The Policy Triangle. Fig. 2. Two- and Multi-sidedness Versus Resellers and VI Firms.  Fig. 3. The Control and Cost Trade-off.  Fig. 4. Heuristic Conceptual Mapping of the Sharing Economy.  Fig. 5. Heuristic Typology. Fig. 6. Renting/Selling, Doing Work, Doing Both. Fig. 7. Renting/Selling, Doing Work, Doing Both by Employment Status. Fig. 8. Renting/Selling, Doing Work, Doing Both: Self-employed and Full-time Employees. Fig. 9. Determining the Role of Science in Policy and Politics.

14 22 33 43 77 98 98 99 165

Tables Table 1. Factors affecting platforms’ size. Table 2. Car sharing versus ride services. Table 3. Selected Litigation Cases in the United States. Table 4. List of Participants.

24 27 112 132

Boxes Box 1. Airbnb Self-reported Impacts. Box 2. Uber’s Self-reported Impacts. Box 3. Conflicts, Bans and Court Cases. Box 4. Online Micro-tasking Ideal-typical Functioning. Box 5. Online Tasking, Ideal-typical Functioning. Box 6. MLM Physical Services’ Ideal-typical Functioning. Box 7. Working Conditions (Investigative Journalistic Accounts). Box 8. Changing Practices by Digital Labour Markets in the United States.

53 53 69 78 80 82 92 112

This page intentionally left blank

Introduction We are being afflicted with a new disease of which some readers may not yet have heard the name, but of which they will hear a great deal in the years to come – namely, technological unemployment. This means unemployment due to our dis’covery of means of economising the use of labour outrunning the pace at which we can find new uses for labour. – John Maynard Keynes, ‘Economic Possibilities for Our Grandchildren’ (1963 [1930]). This book tackles head on and deconstructs the rhetoric surrounding the socalled sharing economy, which has obfuscated so far the public debate, contributing to delays, or total lack of policy and regulatory intervention. In a nutshell, we show that platforms have managed to enlist in their communication and lobbying activity two apparently contradictory rhetorics: that of disruptive innovation typical of neo-liberalism ideology together with the positive normative discourse about grass-roots and bottom-up revival of community, trust, social capital, and the moral economy allegedly enacted by exchanges between peers. Economist Timothy Taylor (2015) argues that the use of the ‘sharing economy’ to refer to various commercial platforms is a ‘triumph of public relations artistry’. Taylor would rather use the expression ‘matching economy’, whereas when we analyse at face value ‘sharing economy’ discourses and selfdefining practices here, we use expressions such ‘platform economy’ or two and multi-sided market. The platform economy has reached sizeable dimensions, and today it is a very politically relevant and hotly debated topic. According to a study released in 2017 by European Commission (EC) (Hausemer et al., 2017), for instance, as many as about 200 million European citizens have used peer-to-peer (P2P) platforms between May 2015 and May 2016. The platform economy is discussed as both source of innovation/growth and matter of various policy concerns going from competition, tax collection, consumers’ protection, privacy and algorithms transparency all the way to the future of work. One particular concern, in fact, regards the alleged advent of technological unemployment predicted in the quotation from Keynes placed at the beginning of this Introduction.

Platform Economics: Rhetoric and Reality in the “Sharing Economy” Digital Activism and Society, 1–15 Copyright © 2019 by Emerald Publishing Limited All rights of reproduction in any form reserved doi:10.1108/978-1-78743-809-520181001

2    Platform Economics The rise of online labour platforms is occurring at the same time as many gurus predict the widespread computerisation and robotisation of jobs, as well as the emergence of a new algorithm-dominated form of governance and organisation of work. Within the broadly defined phenomenon of online platforms, at face value the ‘sharing economy’ seemed a new and refreshing movement where the passions and the interests could be reconciled along the ways suggested by Albert Hirschman (1977, 1982), who, in his attempt to rebuke the pessimism implicit in Mancur Olson’s (1971 [1965]) work on collective action, had argued that social and publicly inspired passions could be reconciled with economic selfinterest. Yet, today the ‘sharing economy’ is contested and critics actually claim that it represents a further encroachment of Neoliberalism economic self-interest on society, rather that the re-embedding of economy within community. Recently, on the basis of 120 in-depth interviews with participants, a group of sociologists suggest that the ‘sharing economy’ can still domesticate the market and build a ‘moral economy’ (Fitzmaurice et al., 2018). At least, these authors claim, most participants see the sharing economy as an opportunity to construct communitarian and morally attuned personalised exchange. Leaving aside the issue of the validity of claims based only on 120 interviews, it is one of our contentions in this book that this kind of narratives are exactly among the sources of the rhetorical discourses obfuscating the debate and used instrumentally by the commercial platforms. Actually, we prefer the way the economic literature deals with both sharing platforms and other platforms, namely as two-sided markets (see infra). At any rate, the developments since 2014 justify the use of the metaphor ‘sharing wars’ (Rauch & Schleicher, 2015) as more appropriate than that of the reconciliation between the passions and the interests to characterise the current debate. The purposive sampling of media and blogs coverage performed for this book, corroborated by other similar exercises (Cohen & Muñoz, 2015; Dredge & Gyimóthy, 2015; Martin, 2016; Richardson, 2015; Schor, 2014; Schor & Attwood-Charles, 2017), indicates that attention peaked in 2014–2015 and increasingly started to focus on controversies and conflictual aspects. The conflict between the ‘passions’ and the ‘interests’ is evident in the claim that ‘true’ and ‘authentic’ sharing and collaborative movements have been hijacked and co-opted in the rhetoric and public relations campaign of big commercial platforms such as Uber and Airbnb to pursue their economic self-interest through traditional lobbying strategies (Caldararo, 2014; Kuttner, 2013; Lee, 2015; Walker, 2015). In less than five years, the ‘honeymoon’ with the ‘sharing economy’ has ended. Optimistic and utopian narratives have been substituted by accounts of legal disputes and of the ‘dark side of sharing economy’ (Malhotra & Van Alstyne, 2014). Therefore, starting especially in 2016, the first sign of policy and regulatory responses occurred in a domain already characterised by heated debates and open conflicts. In the digital single market strategy unveiled in 2015, the EC announced an indepth review of platforms (European Commission, 2015a, 2015b), which translated in a number of initiatives, including, among other things, the e-commerce

Introduction    3 package,1 the communication on online platform (European Commission, 2016c), and the communication on the collaborative economy (European Commission, 2016b); several European countries also have taken action in this domain.2 Yet, as of today, online platforms still function somehow in a grey area not fully governed by the rule of law. In a speech pronounced on 28 February 2018, the President of the European Parliament Antonio Tajani affirmed that currently online platforms are to a large extent legibus solutus (exempt from law), often with dominant market position and paying little taxes.3 In particular, Tajani singled out the ‘collaborative economy’ (official EC policy jargon for what in the majority of academic contributions is called the ‘sharing economy’) and made the example of Airbnb allegedly exempted (according to Tajani) from norms on workers, consumers, taxes, security and licence that on the contrary its microenterprise competitors in the tourism industry must respect. Tajani concludes that the message of the European Parliament is clear: The EU market, besides promoting innovation, must also guarantee competition rules and equal conditions for all players. In practice, the situation in Europe today is not as clear-cut as in the words of the president of the European Parliament, since different ‘regulatory’ solutions are emerging in different countries as a patchwork of judges’ decisions and local authorities’ initiatives. In certain European cities, for instance, Airbnb can operate only at the condition of withdrawing due taxes whereas in others strict permissions and licences are required for those who want to put their house on the platform.4 Yet, there is not a common and systematic common European approach and at times not even a common national-level framework (i.e. in Spain, 1

The e-commerce package contains, among other things, the legislative proposal 2016/0152 (COD) on geo-blocking and the legislative proposal 2016/0148 (COD) concerning unfair commercial practices in the digital world. 2 For instance: (a) In 2015, the French government promoted the Digital Republic Bill, with the objectives of opening up data and knowledge dissemination, ensuring equal rights for internet users, and promoting an inclusive digital society (http://www. republique-numerique.fr/pages/digital-republic-bill-rationale). (b) In 2016, the Italian parliament drafted the Sharing Economy Act (http://www.camera.it/_dati/leg17/ lavori/stampati/pdf/17PDL0039770.pdf). (c) The Dutch Ministry of Economic Affairs commissioned a study to identify and evaluate policy options for online platforms (https://www.tno.nl/en/about-tno/news/2016/3/how-policy-makers-can-deal-withdigital-platforms/. (d) In Germany, the Ministry of Economics has published a Green Book on Digital Platforms outlining rules and framework conditions for online platforms, while the competition authority has carried out an analysis of online market structures. 3 Opening speech by EP President Antonio Tajani at EU-China Tourism Year Parliamentary Day, Brussels, Belgium, 28 February 2018 (see: http://www.europarl.europa.eu/thepresident/it/sala-stampa/intervento-di-apertura-evento-anno-del-turismo-ue-cina). 4 Three impulse chapters requested by EC on how Airbnb operates in different European cities provide an exhaustive picture of this fragmentation of approaches (Ranchordás, 2016; Rating Legis, 2016; Smorto, 2016). The most recent overview of regulatory approaches in the tourism sector for the sharing economy can be found in a report published in 2018 by the EC (VVA & Spark, 2018).

4    Platform Economics rules on Airbnb differs between Madrid and Barcelona; see infra). Whereas various regulations and licence schemes can be applied to Airbnb and other platforms in the accommodation businesses, only two member states/cities (France and Italy) have introduced specific regulatory regimes addressing online collaborative platforms in the accommodation sector. On the other hand, Catalonia has introduced a regulatory regime, which, although not explicitly mentioning online platforms, targets them (and differentiates Barcelona from Madrid). A similar situation characterises platforms involving labour matters and the much-debated issue of whether Uber’s drivers or Foodora’s riders are selfemployed contractors or de facto employees.5 On such labour matters, governments have so far remained idle, and let these be decided by courts. Courts are deciding differently in different countries. In October 2016, a London employment tribunal upheld the claim brought by two Uber drivers on behalf of a group 19 Uber workers who argued that they were employed rather than working for themselves.6 In Turin, Italy, on 11 April 2018, the court rejected an appeal by six riders of Foodora, who, after having participated in a mobilisation for their rights, were terminated by the platform.7 The judges accepted the platform’s claim that the riders were self-employed and could be terminated any moment, and thus could not have the benefit of the right to self-organise as would be done by employees. No doubt that the emergence of online platforms is a case in point of tension and friction between fast technological innovation and regulation where it is more difficult that technology and regulation co-evolve to ensure both innovation and a fair management of risks and sharing of benefits (Brownsword & Goodwin, 2012; Brownsword & Somsen, 2009). In the literature on law and technology, this situation is called ‘regulatory disconnect’, whereby regulation cannot evolve as fast as its target and becomes disconnected (Butenko & Larouche, 2015). Some of the platforms of the sharing economy, such as, for instance, Airbnb, further complicate the picture as they operate at the intersection between different levels of law and regulation: general national law, civil codes provisions, city regulations, and regional-level safety frameworks for rental and in specific rental laws. Yet, there is also no doubt that online platforms have followed a ‘fait accompli strategy’, reached scale and power and then started to deal with compliance only when forced (Degryse, 2016, p. 15). In this way, platforms first avoided existing laws and regulations, which enabled them to later claim that they are a specific case and should be either exempted from existing regulatory framework or be the

5

For the most updated overview of policy and regulatory developments concerning the issue of workers of the sharing economy in Europe, see a report recently published by the European Agency for Safety and Health at Work (Garben, 2017.). Other two reports also published by the Commission provide an overview of regulatory developments in EU28 concerning consumer protection (Hausemer et al., 2017) and of cases concerning sharing economy platforms that have been dealt by the Court of Justice of the European Union (Psaila, Fiorentini, Santos Silva, & Gomez, 2017). 6 Aslam & Ors v. Uber BV & Ors [2016] EW Misc B68 (ET) 28 October 2016. 7 See, for instance, Cruciatti (2018).

Introduction    5 target of innovative regulations. Second, this enabled to gain wealth and influence and the needed political power. Third, they first locked in consumers and workers to become reliant and dependent on them and now act as their political supporters. Hence, it is debatable, to what extent their success is attributable to pure innovation or rather to ‘regulatory arbitrage’.

A Rhetoric-Driven ‘Negative Policy Bubble’ It is the claim of this book that such strategy of ‘fait accompli’ was not only the result of regulatory difficulty and platforms’ ability but it is rather a case of a ‘negative policy bubble’ skilfully obtained through rhetorical framing and lobbying. The concept of ‘policy bubble’ has been put forward to provide alternative explanations to when we observe an ‘oversupply’ or ‘undersupply’ of policy responses (Jones, Thomas, & Wolfe, 2014; Maor, 2014, 2016). For what concerns the systematic undersupply of policy, common explanations include, among others, institutional frictions (Baumgartner et al., 2009; Jones & Baumgartner, 2005), policy drift due to pressure from veto points in the political process (Hacker, 2004) and time trade-offs over time when government is unwilling to impose immediate costs for future gains (Jacobs, 2011). The ‘policy bubble’ perspective introduces a cognitive and emotional dimension of analysis related to the heuristics and biases that can be opportunely framed through rhetorical narratives, activated by those interested in producing an oversupply or undersupply of policy responses. Since the emergence of the first ‘sharing economy’ platforms in Europe in the mid of the first decade of the twenty-first century until at least 2016, there has clearly been an undersupply of systematic policy and regulatory responses, testified by the earlier cited February 2018 speech of the president of the European Parliament basically stating that they still function to a large extent outside the rule of law. After platforms started operating and reached scale, they have skilfully enacted lobbying strategy based on rhetorical framing and on the instrumental and selective use of empirical evidence. Cannon and Summers (2014) wrote a piece providing strategic suggestions to Uber and other ‘sharing economy’ companies on how to fence current criticisms and win over ‘regulators’ by using state of the art techniques to reach out to government, produce well-researched case showing the benefits created by companies, use external validators and create coalitions. Indeed, all the major commercial sharing platforms have just done exactly as suggested. The narrative of their blogs is clearly framed in terms of sharing and caring and have produced several self-reports on the many economic and social impacts they allegedly deliver to economy and society (Codagnone, Abadie, & Biagi, 2016b). Just a few examples are reported here. In a public hearing in the UK House of Lords (2016), Patrick Robinson (Head of Public Policy Europe and Canada for Airbnb) affirmed: In our case, the public interest at stake here is, first, about consumers and consumer choice not just to consume services but to be producers of services too. The additional income that Airbnb

6    Platform Economics hosts are making is very important to them. Identifying outdated rules and regulations that might stop people engaging in what is beneficial activity is a good exercise and one that I am delighted that we undertook in London. The rhetoric of a flat world is integrated with that of the advent of a global online meritocracy by the Elance-oDesk (2014, then renamed Upwork) annual impact report. On 10 February 2016, Airbnb, Uber and 45 other commercial ‘sharing’ platforms sent an open letter to the Netherlands Presidency of the Council of the European Union,8 where they demanded to be protected from regulatory intervention taken at national and local levels in view of their great contributions to sustainable economic growth in Europe. This book shows that evidence is emerging on both positive and negative effects of platforms, but it also unequivocally documents that the currently available evidence on costs and benefits is absolutely partial, not yet conclusive, and does not warrant the above claim through which platforms ask protection of the EC in view of possible threats from national- and local-level regulatory interventions. Not surprisingly, several observers have claimed that large companies have co-opted the sharing movement to pursue economic self-interest through traditional lobbying strategies (Lee, 2015; Schor, 2015; Walker, 2015). According to Lee (2015, p. 17), the ‘sharing economy’ is just another example of how ‘insurgent sentiments’ are used to ‘sell the bona fide of profit-making corporations’. The anti-establishment ideology disseminated by magazine Sharable and associated Peer.org is increasingly seen as a mouthpiece of big companies, such as Uber and Airbnb, which use such rhetorical weaponry for the pursuit of their economic interests (Kerr, 2014). As a matter of fact, it is our claim that the sharing economy is a unique case in which concentrated and specific economic interests (large commercial platforms) not only have used to their advantage the narratives describing the initial phase of the sharing movement but they have also succeeded in the infrequent objective of enlisting diffuse (i.e. consumers) interests as their allies. This book aims precisely at tracing the origin of such rhetorical framing and at deconstructing them by way of analysing the available empirical evidence around the underlying claims. In this, we follow the inspiration from the general approach of Hirschman (1991), and particularly from his book The Rhetoric of Reaction. Throughout his scientific production, Hirschman considered ideas, values and rhetorical discourses as having autonomous effects on the process of change itself, regardless of whether or not they are empirically grounded. He considered them part of the endogenous mechanisms of social and economic change with an approach that can be deemed ‘pragmatic idealism’ (Adelman, 2013, p. 422). In The Rhetoric of Reaction, Hirschman observed how opposing groups in liberal democracies sometimes get walled off from each other’s opinions and views; rhetorical discourses can explode into conflict simply as a result of the ‘imperative of

8

Retrieved from https://www.airbnbaction.com/wp-content/uploads/2016/02/NLCouncilLetterCollabEcon-Final-100216-4.pdf (accessed on 9 October 2017).

Introduction    7 the argument’. Rhetorical discourses limit what people might consider as alternatives, are immune from being wrong and accommodate uncertainty. Hirschman considered instead a detached analysis of surface rhetoric, placed historically and analytically in context, more useful than a head-on attack on one of the opposing factions, and claimed that deconstructing rhetoric by using empirical evidence could help restore dialogue and communication between conflicting factions. Rhetorics are part and parcel of debates on important policy issues that involved opposing interests entering into various forms of negotiations that can be settled or become intractable. Rhetorical discourses do not emerge from nowhere but are historically inspired and recurring. Hirschman compared, for instance, the neoconservative attacks on welfare states, such as Charles Murray’s Losing Ground (1984), with the reactions of hundreds of years earlier against the ‘poor laws’. He noted how ‘any idea that has been out of view for a long time has a good chance of being mistaken for an original insight’ (Hirschman, 1991, pp. 29–30). A case in point is the discourse about gig workers performing tasks on digital labour markets for ‘pin money’. Here, an old idea first articulated in the 1950s and 1960s about the then-emerging temporary work agencies in the United States has clearly resurfaced (Berg, 2016). In sum, the ‘sharing economy’ is today a rhetorical field that needs unpacking, and this book contributes to this undertaking. It may seem churlish to deconstruct these discourses with empirical evidence and to challenge claims made by both naive disinterested and shrewd self-interested parties about le magnifiche sorti e progressive (the magnificent and progressive fate) of the ‘sharing economy’. Alternatively, deflating the gloomy predictions of the harshest detractors of ‘sharing’ platforms may be considered apologetic. Yet, this is exactly what this book aims to do.

Elective Affinity There is a clear elective affinity between the first narratives on the sharing economy, the conceptual ambiguity they have generated, and the way this has been harnessed into rhetorically framed discourses. The expression ‘sharing economy’ is actually a ‘floating signifier’ for a diverse range of activities (Nadeem, 2015, p. 13), and its usage to refer to various commercial platforms can be seen as the ‘triumph of public relations artistry’ (Taylor, 2015). As discussed by Belk (2014a), among all the platforms included into the sharing economy one could distinguish ‘real sharing’ from ‘pseudo sharing’, by which he means ‘business relationship masquerading as communal sharing’. The conceptual ambiguity is such that at times one is left to wonder whether stakeholders, experts and policy makers are talking about the same phenomenon. It is important to note that there is a closed self-reproducing loop between conceptual ambiguity, rhetorical controversies, lack of sound measurements and empirical evidence, and fragmented or non-existent policy and regulatory approaches. Such conceptual ambiguity goes hand in hand with rhetorical narratives that have been harnessed into current disputes between boosters promising ‘utopian outcomes’ (empower consumers, lower carbon footprint and efficiency) and the

8    Platform Economics critics denouncing that self-defined sharing companies are in reality about economic self-interest pursued in predatory and exploitative manners. Under the label ‘sharing economy’, depending on the stream of literature, one can find all sorts of different platforms that have in common only the fact that they have two groups, that is, users and providers, increasing the scale and speed of traditional transactions such as selling, renting, lending, labour trading and provision of services. Such usage is confusing to the point of making expressions such ‘sharing economy’ or ‘collaborative economy’ conceptually trivial. The attempt to place or be placed within such label is part of what can be called the rhetorical politics of platformisation. Also, the digital labour markets, which we analyse in this book, and include both that allow the remote delivery of electronically transmittable services (i.e. Amazon Mechanical Turk, Upwork, Freelancers, etc.) and those where the matching and administration processes are digital but the delivery of the services is physical and requires direct interaction, are at times referred to as sharing platforms. In practice, however, sharing platforms diverge in terms of dimensional relevance (i.e. from a few hundred users to millions of users), interaction modality (i.e. P2P vs business-to-consumers) and type of assets being exchanged (i.e. a property vs one’s labour). Hence, they differ also in terms of their current policy and regulatory implications (e.g. market access and licensing, liability and insurance, consumer protection and labour laws). As this book deals with rhetorics and such rhetorics are closely connected to such conceptual ambiguity, we will use expressions such as ‘sharing economy’ or ‘sharing platforms’, although in practice we believe that the most appropriate way to conceptualise and treat them is that used in economics where these players are discussed as two-sided markets or two-sided platforms. Since 2002, a growing body of mostly conceptual–theoretical economic literature (discussed in Chapter 1) has analysed situations where one economic operator (originally referred to as an intermediary and later increasingly as a platform) brings together at least two different groups of users as instances of ‘two-sided’ or ‘multisided’ (when there are more than two groups) markets. Yet, narratives and definitions of the sharing economy have emerged with emphasis on social and communitarian elements, which have proven to be very sticky and also rhetorically useful, especially for those commercial platforms that have scaled up. This has created a paradox (Richardson, 2015) and a built-in potential for polarisation (Martin, 2016). The paradox consists of the fact that the sharing economy is defined simultaneously as part of the capitalist economy and as an alternative to it. This resulted in the polarisation of sharing economy as a new socially and environmentally sustainable form of consumption/production or as a new nightmarish form of Neoliberalism. In their analysis of discourses on the ‘sharing economy’ and its impacts, Dredge and Gyimóthy (2015, pp. 2–3) showed that the initial framing of issues publicly debated has created path dependencies that prioritise certain aspects to the detriment of others and end up determining the agenda for public and policy discourses and debates. The roots of rhetorical framing can be found in the optimism that accompanied the initial phase of ‘crowds’ and ‘sharing’ phenomena. First came the optimism of crowdsourcing and that of the ‘wisdom of crowds’ (Anderson, 2006;

Introduction    9 Benkler, 2004, 2006; Benkler & Nissenbaum, 2006; Surowiecki, 2004), allegedly offering a ‘cognitive surplus’ (Shirky, 2010) and problem-solving capabilities (Brabham, 2008, 2013; Gehl, 2011) and promising new efficiencies (Chandler & Kapelner, 2013; Djelassi & Decoopman, 2013; Satzger, Psaier, Schall, & Dustdar, 2013). Then came the social optimism on the sharing movement portrayed as offering triple wins: greener commerce, greater profits and rich social experiences in the form of community revival and strengthening of social capital (Grassmuck, 2012a; Leadbeater, 2009; O’Regan, 2009; Wittel, 2011). Later on, management (Guttentag, 2013; Heimans & Timms, 2014; Matzler & Kathan, 2015; World Economic Forum (WEF), 2013, 2014; Wosskow, 2014) and neo-liberal economics’ optimistic narratives followed (Allen & Berg, 2014; Cohen & Sundararajan, 2015; Koopman, Mitchell, & Thierer, 2014, 2015; Sundararajan, 2014; Thierer, Koopman, Hobson, & Kuiper, 2015). Management gurus (Heimans & Timms, 2014), for instance, proposed a distinction between ‘new power’ (sharing economy but also grassroots political movements) and ‘old power’ (big corporations, but also established political parties) where the former is about radical transparency, openness and collaboration, and the latter about bureaucracy. Neoliberal and libertarian economists expect the ‘sharing’ platforms to: (a) increase economic activities and productivity through better use of underutilised assets or ‘dead capital’ and through lowering transaction costs that expand trade; (b) increase social utility and consumer welfare as a result of more competition; (c) create new jobs; (d) reduce information asymmetry between consumers and producers, thanks to reputational ratings; (e) create new markets through disruptive innovations and spur in turn further innovation among incumbent industries; and (f) produce a new cohort of entrepreneurs if the micro-entrepreneurs providing services to the platforms acquire the experience and skills to progress and launch their own ventures. In sum, conceptual ambiguity and rhetorical discourses fuel controversies between supporters and opponents, who harness and present conflicting ad hoc ‘evidence’. Disputes flourish, as there is no basis to adjudicate opposing claims. The practice of platforms not to disclose important metrics, or to make them available only to some researchers, further contributes to a debate that is not informed by evidence. Hence, not only in the press and in reports by politically positioned think tanks but also in many peer-reviewed academic essays one finds value-loaded, normative and prescriptive claims.

Interpretative Framework: Sharing Rhetorics and Evidence-Based Policy The instrumental rhetorical discourses, when empirical evidence is still lacking or inconclusive, have contributed to the undersupply of policy and regulatory responses and today to policy and regulatory decisions taken ‘in the dark’ often under the influence of only some interested groups; one also has the impression that politicians and policy makers have abdicated their role and are mute, while courts and judges pronounce their judgements. And yet, given the growth of sharing platforms, there are several issues at stake for

10    Platform Economics several groups of interest on which decisions are needed, such as: (a) for users as consumers, balancing between the alleged benefits from cheaper and more convenient choices and possible risks (i.e. liability issues, safety, information on quality of products and services, etc.); (b) for users as providers (i.e. the ‘contractors’ performing work), balancing between new employment opportunity and the erosion of ‘labour contract’ and social protection; (c) for platforms that have much to gain or lose depending on regulatory decisions; (d) for established operators (i.e. traditional taxi and traditional accommodation), potentially disrupted by sharing platforms, who have already challenged such platforms in both national courts and the Court of Justice of the European Union (CJEU); and (e) last but not least, for the general public interest, since sharing platforms can have positive or negative externalities on the economy and society as a whole (i.e. positive surplus from innovation versus erosion of tax base). The ‘sharing economy’ qualifies as a domain where values are disputed, facts are uncertain and stakes are high.9 As a matter of fact, it is our claim that the ‘sharing economy’ is one among the many policy domains that are bringing to the fore the intrinsic limitations and crisis of rudely positivistic and technocratic nature of the evidence-based policy (EBP) paradigm.10 While there is hardly anything new in the programme of using evidence for policy (Bogliacino, Codagnone, & Veltri, 2015; Misuraca, Codagnone, & Rossel, 2013; Pawson & Tilley, 1997; Vedung, 2010), the main peculiarity of the EBP paradigm is its unrealistic ambition to eliminate any ideological element and judgements from the formulation of policies, and to curb the discretion of professionals (i.e. teachers, field workers, policy officers, etc.). Indeed, it has been criticised as sort of new ‘rationality project’ to expel politics from policy-making that consider democratic processes as rent-seeking and a deadweight loss to society (Kay, 2011). Increasing criticisms of EBP include authors turning it on its head and arguing that what is happening in practice is ‘policy-based evidence making’ (PBEM) as a form of misuse of evidence in policy-making (Sanderson, 2011; Strassheim & Kettunen, 2014; Torriti, 2010). That policy can be entirely determined by evidence is the naïve and technocratic tenet of the EBP paradigm. In fact, we see policy as shaped by the interaction between three dimensions: evidence, interest groups’ politics and values. This framework has been developed and applied by Codagnone, Bogliacino, and Veltri (2018a, pp. 24–30) and Codagnone (2017) and called it the ‘policy triangle’. Here, we use this interpretative framework for the purpose of discussing the role 9

This characterisation comes from the literature on Science and Technology Studies and in particular from a ‘post-normal science’ approach (Funtowicz & Ravetz, 1990, 1991, 1993, 2008; Ravetz, 1990). 10 The EBP was launched as part of the programme of New Labour in Britain Blair (Cabinet Office, 1999), but rapidly spread beyond United Kingdom (Nutley, Morton, Jung, & Boaz, 2010); as noted (Head, 2013, p. 397), most policy makers adhered swiftly to the EBP mantra, since policy-making based on ignorance, opportunism and vested interests is not ‘readily admitted’. For more detailed discussion on the emergence and crisis of EBP, see Codagnone et al. (2018a, chapter 1).

Introduction    11 of rhetorical framing within the interaction between policy makers and interest groups with respect to the sharing economy. The evidence dimension concerns science and scientists and how they influence, or are influenced by, policy-making debates and the values of the surrounding society (Pielke, 2007). The values dimension, with its underlying emotionally shaped system of belief,11 in a stylised fashion refers to society and the citizenry at large. The dimension of politics has to do with both organised interests and policy makers, and in the following section we concentrate on it. The political dimension of policy-making involve also the policy-making bodies and the policy makers because it would be naïve to take for granted that policy is enacted only for the public interest and that evidence is used in policymaking only for the sake of efficiency and effectiveness. Policy makers have their own agenda and goals as well as their values.12 They interact with concentrated specific interests (i.e. industry) and diffuse interests (i.e. consumers). This is a classical distinction following Mancur Olson’s (1971 [1965]) theory of collective action. Olson (1971 [1965], p. 166) deemed diffuse interests as the ‘forgotten groups’ and took consumers as a typifying example of a numerous but dispersed group that ‘have no organization to countervail the power

11

For social psychological discussion on the concept of values we follow the definition provided by Thorngate (2001, pp. 88–91). An attitude is an attraction/repulsion to/from a thing, idea, concept and person, which is called the attitude object (i.e. a policy or Attitude Object [ATO]). The attraction/repulsion is felt as an emotion. The degree of attraction/ repulsion is called a value. When we say that a policy has a value, it means we have an emotional reaction to it, or an attitude. Often we have a set of emotional reactions that can be combined into an overall attitude. Each item in this set, however, reflects a link between our emotion and belief. 12 According to classical Public Interest Theory, policy pursues common and public goods, and addresses market failures. As first advanced by Pigou (1920, 1932), regulation is supplied in response to the demand of the public for the correction of inefficient or inequitable market practices. Regulation is assumed initially to benefit society as a whole rather than particular vested interests. The rival of the Public Interest Theory has been called the ‘Capture Theory’ (Posner, 1974) or alternatively the ‘Chicago Theory’ (Hantke-Domas, 2003, p. 165). The first contribution in the construction of this rival approach came from George Stigler (1971), who wrote about the influence of interest groups in designing and enforcing regulation, arguing that regulatory agencies may be ‘captured’ by special interests. From an organisational and institutional perspective, it has also been shown that policy-making bodies as any other organisation do not pursue only instrumental goals but also their own survival and legitimacy (DiMaggio & Powell, 1983, 1991; Selznick, 1948, 1949). Finally, in a groundbreaking and neglected social psychological analysis of the role of policy analysis, Thorngate (2001) sheds light on the social-psychological deviations from instrumental rationality in policy-making. His main argument is that policy makers decide to make or improve policies, not only for the intrinsic merits of the policy at stake but also for competition with peers, to get promotion or avoid demotion, for the desire to save face and that often policy-making is ridden with social influence and group dynamics distortion (i.e. normative conformance to avoid embarrassment, censure or ostracism, and group processes such as group thinking bias).

12    Platform Economics of organized or monopolistic producers’.13 Specific and concentrate interests (they being an association or a single firm) have a better capacity to access and/ or buy the evidence to be exchanged with policy makers. Evidence, in fact, has become the new currency of lobbying to influence policy-making and framing policy debates. Two books written in parallel, one on Washington (Drutman, 2016) and another on Brussels (Laurens, 2017) lobbyists, have shown in detail how evidence is strategic for lobbying and is a source of power asymmetry between the interests that have money to buy evidence and those who do not. It has been estimated, for instance, that in the United States for each dollar spent by the trade unions to produce evidence, corporations spend $34. Many reports on impacts self-published or commissioned to high profile scholars or to former members of executive power by Airbnb and Uber (discussed in Chapter 2) are case in point of this aspect and seem to apply the strategic advice contained in the earlier cited piece of Harvard Business Review (Cannon & Summers, 2014). It is in such a context that the issue of rhetorical framing acquires importance as an instrument within the interest groups’ politics. Rhetorical discourse is an instrument of framing policy agenda and debates. Tversky and Kahneman (1981) have shown that framing can affect the decisions taken in any given choice problems. So, a framing strategy can be used by players on both sides of a policy-contested issue to polarise the situation. The importance of framing was first adopted in the sociological analysis of social movements (Benford & Snow, 2000; Klandermans, 1997; Snow & Benford, 1988; Snow, Rochford, Worden, & Benford, 1986). Frame alignment, bridging and amplification are processes of strategic importance in the mobilisation of social movements (Snow & Benford, 1988; Snow et al., 1986). Frame alignment occurs when individual frames are linked and made congruent, and thus produces ‘frame resonance’ and catalyses the group-formation process. Frame bridging involves the ‘linkage of two or more ideologically congruent but structurally unconnected frames regarding a particular issue or problem’ (Snow et al., 1986, p. 467). Frame amplification refers to ‘the clarification and invigoration of an interpretive frame that bears on a particular issue, problem, or set of events’ (Snow et al., 1986, p. 469). Through the value amplification 13

According to an extensive review of interest groups’ politics (Beyers, Eising, & Maloney, 2008, p. 1109), the distinction between specific and diffuse interests remains quite influential, but it is rather an empirical question to be tested. Bouwen (2002, 2004), analysing the multitude of access opportunities used by groups to exert influence in the EU’s multi-levels, stresses upon the importance of such groups to provide EU institutions (i.e. commission, council and parliament) with the ‘access goods’ (basically information) that they demand. Bouwen (2002, p. 370) defines access goods as goods provided by private actors to the EU institutions in order to gain access. Each access good concerns a specific kind of information that is important in the EU decision-making process. The criticality of an access good for the functioning of an EU institution determines the degree of access that the institution will grant to the private interest representatives.

Introduction    13 logic, framers can actively promote and embellish a specific value to justify the actions proposed in its name. Value amplification refers to the identification, idealization, and elevation of one or more values presumed basic to prospective constituents but which have not inspired collective action for any number of reasons. (Snow et al., 1986, p. 469) Following these insights, in the last two decades the framing perspective and the role of ideas have increasingly been applied both theoretically and empirically in the study of both politics and policy-making. According to Béland (2009), framing affect the policy-making process in the following three ways: (a) constructing the issues entering the agenda; (b) shaping the assumptions that affects the content of policy proposals; and (c) it may build discursive weapons in the construction of reforms imperatives. Particularly important is the role of framing in shaping policy assumptions, as they form what Hall (1993) calls ‘policy paradigms’. Therefore, in our policy triangle framework, the rhetorical framing of discourses, occurring also by way of selectively producing and disseminating ‘evidence’, can be considered a strategic weapon of lobbying, especially by concentrated interests in an attempt to shape policy agenda and contents in ways that are favourable to them or at least to contain the potential damage that may derive from regulatory intervention. Rhetorical framing has been a key weapon used by the most concentrated and powerful economic interests, which are active in the sharing economy policy and regulatory arena. As we shall see in more detail in Chapters 2 and 3, the following rhetorical framings emerged on the sharing economy in general: (a) revival of community and increase in social capital; (b) more equitable distributional effects; (c) positive environmental and economic impacts; (d) disruptive innovations; and (e) capacity for self-regulation through peer reviews’ normative control (i.e. ratings). With regard to online platforms’ intermediating labour, there are at least four specific rhetorical framings. First, the rhetoric of a flat world allowing digital labour migration with no boundaries and a world online meritocracy. Second, the discourse on extra money as a motivation to work for flexors (students, retirees, stay at home parents, etc.). As acutely observed by Berg (2016, p. 18), the claim about individuals working in digital labour markets for ‘pin money’ or out of boredom is a replication of the rhetoric used in the late 1950s and in the 1960s when the new temporary agency industry in the United States was portrayed as employing just middle-class wives killing time and earning extra money. This is an emblematic case of Hirschman’s claim that rhetorical discourses of the past tend to resurface. Third, the alleged contribution of online labour platform to bring back to work the unemployed and underemployed. Fourth, discourse on flexibility, autonomy and creativity that these platforms allegedly provide to their workers. The present book maps the rhetorical framings that shape policy debate on sharing platforms against the available empirical evidence in order to reduce the obfuscation effect and possibly favour more evidence informed policy and regulatory decisions. Going back to Fig. 1, our aim is to help the debate move to the kernel of evidence-informed pluralistic policies. Currently, the rhetorical framing

14    Platform Economics

Interest groups polics

Policy Based Evidence making Evidence informed pluralistic policies

Evidence

Values

Fig. 1:  The Policy Triangle. Source: Codagnone (2017, p. 19). and the disputes are instead producing a mix of the other three types identified in the figure: PBEM, issue advocacy and post-truth. This does not mean that we aim to provide a new technocratic EBP solution. The book unpacks the rhetoric, removes semantic and conceptual confusion, and identifies what empirical evidence is available and what is missing. The principle of scientific analysis based on sound design, methods and evidence gathering is not abandoned in favour of a relativistic and constructivist account. On the contrary, the book is strongly rooted in the scientific method, and starts from a ‘humble’ premise that scientific research will not solve all disputes. It follows that sources were selected and analysed to capture both empirical evidence and rhetorical discourses.

Sources and Structure of the Book The work that went into this book started in May 2015 and was completed in May 2018, thus involving three years of research based on secondary sources, firsthand in-depth analysis of platforms and the field work (presented in Chapter 4). Although earlier versions of this work have been presented in other publications,14

14

A general analysis of the sharing economy (Codagnone et al., 2016b); a scoping exercise on the sharing economy that placed it in the context of the literature on twosided market (Codagnone & Martens, 2016c); a more specific analysis of digital labour platforms(Codagnone, Abadie, & Biagi, 2016a); a first and preliminary essay on the sharing economy as a form of rhetorical framing (Codagnone, 2017).

Introduction    15 the account presented in this book is original in at least three ways. First, the evidence base has been updated to mid-2018, whereas previous versions were based on sources available as of mid-2016. Second, a new and fully grounded theoretical and conceptual interpretative framework has been applied. Third, a case study based on in-depth fieldwork in three European cities examining the ideological production in digital intermediation platforms has been added. This book is one of a kind because it gathers extensive primary and secondary evidences (a total of about 864 unique sources) and takes an inter-disciplinary approach in which economics, sociology, anthropology, legal studies and rhetorical analysis converge. These sources included (a) extensive analysis of media accounts and blogs (about 220 sources); (b) formally reviewed academic literature on sharing and labour platforms (about 234 sources); (c) other indirectly and contextually relevant academic literature (about 200 sources ranging from the economics of two-sided market to the study of non-standard work, computerisation and the future of work, social capital, analysis of ratings, etc.); (d) grey literature (about 100 items, including both policy reports and reports by interested parties such as Airbnb, Uber, other platforms and trade unions); and (e) in-depth review of a purposive sample of platforms (110 in total, of which 70 on renting/selling and 40 on intermediating labour-based services).15 The fieldwork in Chapter 4, which analyses key strands of rhetoric in platform ideologies, includes 28 in-depth interviews of sharing economy and alternative governance players in Barcelona, Paris and Berlin during November 2015 and February 2017. Given the diversity of sources and the objectives of this book, the narrative alternates theoretical and conceptual reasoning, discussion of ‘hard’ findings from experimental and quasi-experimental studies, and ‘softer’ issues such as rhetorical discourses and media-‘hyped’ accounts. Such accounts are contrasted with the limited empirical evidence available on key aspects (motivations to participate, trust and social capital, platform-matching and rating mechanisms) and impacts (environmental, economic and social). In Chapter 1, we present an introduction to economic literature on two-sided market as the background to contextualise commercial sharing platforms and as a basis to unpack rhetorical discourses and anticipate some of the key policy and regulatory issues. Chapter 2 deals with the sharing economy in general by presenting a typology, describing the main rhetorical framing and contrasting them with empirical evidence. Chapter 3 repeats the exercise of Chapter 2 but with an in-depth zoom on digital labour platforms. Chapter 4 presents an in-depth discussion and analysis of three strands of ideological discourse (neoliberal, commons and platform co-operativism) in the sharing economy. Chapter 5 concludes with a critical review of open policy and regulatory issues and the corresponding evidence gaps that call for future research.

15

Only a selection of these sources is referenced in the book; the platforms analysis can be found in the annexes of Codagnone et al. (2016a, 2016b).

This page intentionally left blank

Chapter 1

Platform Economics and the Sharing Economy: A Primer Introduction The first question to ask about the ‘sharing economy’, here we use the expression in its widest definition to include also digital labour platforms, is a conceptual one: What is so peculiar about the platforms labelled as ‘sharing’ that sets them apart from other digital platforms that have never been considered as part of the sharing economy? Answering this question would represent the first step towards removing the effects of rhetorical obfuscation. Current definitions of the sharing economy are mostly ‘ostensive’ (by pointing and exemplification) rather than ‘intensional’ (connotative),1 which prevent to build a sound typology that (1) have descriptive power and be empirically grounded, (2) reduce complexity and (3) identify similarities and differences. Typologies are organised system of types that can be used for ‘forming and refining concepts, drawing out underlying dimensions, creating categories for classification and measurement, and sorting cases’ (Collier, LaPorte, & Seawright, 2012). Typology mapping ensures greater parsimony and is instrumental for both theory

1

Intensional definitions are clear-cut in that they establish the necessary and sufficient conditions for a ‘thing’ being a member of a specific set. The advantage of intensional definitions is that they automatically produce mono-dimensional classifications. The disadvantage is that empirical reality is always more complex and nuanced than an intensional definition could capture; using the intensional approach may entail excluding items from a specific set of ‘things’ in ways that may appear artificial or arbitrary, especially when this contrasts with how players self-define themselves or are defined by others in practice. Ostensive definitions more pragmatically denote just a few key features and complement them with exemplifications. The advantages are that they are more inclusive (reducing hard clearcut choices and exclusions) and fairly easy to be produced. The clear disadvantage is that if they are too loose and encompassing, they become trivial with limited descriptive power and do not reduce complexity as they group together entities that are similar regarding very few characteristics (possibly not the most relevant ones) and very dissimilar regarding a much larger number of relevant ones. Platform Economics: Rhetoric and Reality in the “Sharing Economy” Digital Activism and Society, 17–34 Copyright © 2019 by Emerald Publishing Limited All rights of reproduction in any form reserved doi:10.1108/978-1-78743-809-520181003

18    Platform Economics development and policy-making.2 Different types, if identified consistently, may have different regulatory and policy implications and some subset can be by analogy assimilated to other already existing and regulated activities. The first step towards conceptual clarification is to reformulate the above question as follows: Should/could ‘sharing economy’ platforms be analysed as twosided and multi-sided markets? Or do they present features that set them apart from others (digital or analogic) commonly studied as two-side markets or multi-sided platforms (MSPs)? If sharing platforms could unequivocally be defined as two-sided markets, this would warrant the application of both established analytical and theoretical approaches, and the corresponding policy ‘tool box’. As we shall see, whether or not a particular activity qualifies as a two-sided market has relevant implications with respect, for instance, to competition policy (Evans, 2003a; Evans & Noel, 2005; Evans & Schmalensee, 2007; Wright, 2004). The first indirect answer could be garnered by looking at those contributions from the literature on two- and multi-sided markets where some of the most well-known ‘sharing’ platforms (i.e. Uber, Airbnb, oDesk today Upwork and TaskRabbit) are studied or at least cited as examples (Cullen & Farronato, 2015; Farronato & Fradkin, 2015; Fradkin, 2014; Fradkin, Grewal, Holtz, & Pearson, 2015; Hagiu & Wright, 2013, 2015a, 2015b, 2015c; Horton, 2014; Horton & Golden, 2015). Within this literature, the expression ‘sharing economy’ is rarely used and platforms are studied simply as peer-to-peer (P2P) two-sided markets for the exchange of underutilised goods and services. The only special characterisation compared to other instances of two-sided markets concern the P2P dimension. This indirect answer, however, is not sufficient to answer the above question. In this chapter, we first provide a primer on the two-sided market literature based on a selective review of key contributions; then in the end we come back to consider the extent to which at least some of the sharing platforms qualify as two-sided markets and their implications.

In Search of Two- and Multi-sidedness Two-sided or multi-sided markets or platforms are situations where a platform enables two or more groups of users to transact or at least interact in ways that at least one group and usually all groups benefit directly or indirectly from having a growing number of users on the other side(s). As not to repeat multiple expressions all the times, henceforth we simply use the acronym 2SMs (two-sided markets) to generically refer to this type of economic operator (including if they are multi-sided or, if technically, they should be called platforms rather than markets).

2

On the processes of classifying in social sciences and on typologies, see, for instance, Bailey (1994).

Platform Economics and the Sharing Economy    19 Hence, 2SMs internalise for the matched sides the network externalities and reduce transaction costs. Network externality means that the value of a 2SM and the number of transactions increase more than proportionally with the number of participants. The higher the number of participants already on the market, the more others will want to join because it increases consumer choice and boosts the pool of customers for service suppliers. This may trigger competition policy questions, although scaling up to dominance (size) and industry concentration are constrained by several factors (see infra). The relevance of the 2SM literature has to do with competition policy implications that have been addressed by several authors (e.g. Evans, 2003a; Evans & Noel, 2005; Evans & Schmalensee, 2007; Wright, 2004); many traditional axioms of economic analysis that inspire competition policy do not hold and should not be used when markets are two-sided. For instance, pricing to one side below marginal cost is not a predatory behaviour but rather a common profit-maximising strategy in 2SMs. Defining the relevant market for antitrust purposes and looking at only one side can lead to a market definition that is too narrow. Furthermore, network effects can lead to tip towards a single dominant platform (Rysman, 2009). As reviewed in Li (2015, p. 100 and pp. 103–105), the two-sided perspective has been used by European Commission (EC) and the European Union (EU) courts, that is, the General Court (GC) and the European Court of Justice (ECJ), when applying the EU competition law. Hence, the question of which markets are two-sided has become increasingly relevant (Filistrucchi, Geradin, & Van Damme, 2013, p. 36). Since 2002, a growing body of economic literature has analysed situations that broadly qualify as 2SMs, although as we shall see, there is no consensus regarding the conditions for two-sidedness (or multi-sidedness), which remains an empirical matter to be ascertained case by case (Filistrucchi, Geradin, van Damme, & Affeldt, 2014; Filistrucchi et al., 2013). The first to look at firms serving two different types of customers and facing the ‘chicken and egg problem’ were Gawer and Cusumano (2002) and Caillaud and Julien (2003b). These authors referred to ‘intermediary markets’ serving two distinct groups of customers. The expression ‘two-sided market’ was first introduced by Rochet and Tirole (2003, 2006) and was used later by Wright (2004) and Armstrong (2006). In parallel, Evans (2003a, 2003b) used the expression ‘two-sided platforms’ and was one of the first to systematically apply this perspective to what he called the web economy (Evans, 2008a, 2008b, 2009, 2011). On the other hand, Parker and Van Alstyne (2000, 2005) were converging on ‘two-sidedness’ coming from network and information theory, and with Eisenmann, Parker, and Van Alstyne (2006) were the first to talk about ‘two-sided strategies’ rather than ‘markets’. Rysman (2009) also used the expression ‘two-sided strategies’ to convey the idea that there are choices made by agents rather than an imposed endogenous industry structure. Hagiu and Wright also look at multisided platforms as a matter of firms’ strategic choices. Building on the theory of the firm, they frame these choices as a trade-off between ‘being a MSP or a vertical integrated firm’ (Hagiu & Wright, 2015c), or between ‘controlling versus enabling’ (Hagiu & Wright, 2015a). Initially, the focus of this literature was made up of payment systems, auctions, operating systems and media markets. Lately, however, it

20    Platform Economics has been increasingly applied to digital platforms. In particular, digital platforms are discussed in the most recent works by Hagiu and Wright (2013, 2015a, 2015b, 2015c). Some digital platforms are also the object of controversy over whether or not they can be considered two-sided (Li, 2015; Luchetta, 2014). As stated, there are different approaches to identify the conditions of twosidedness, which can be ascertained only empirically by considering specific cases. Following the analysis presented by Li (2015), three approaches are identified regarding how two or more groups of users interact: (1) two groups of customers exert bilateral indirect network externalities; (2) only one group of customer exerts unilateral indirect network externalities on the other; and (3) the existence of indirect network externalities is not necessary. Below, these are discussed briefly together with the more specific perspective of Hagiu and Wright (2015c) on the difference between 2SMs, resellers and vertically integrated (VI) firms. Two-way indirect network effects. In Roche and Tirole’s (2003) original contribution and in the early analysis of intermediary markets by Caillaud and Jullien (2003a), indirect network effects on both sides are fundamental, and the role of the platform (intermediary) is to internalise the externalities produced by the fact that the decision of each set of agents affects the outcomes of the other set of agents (Rysman, 2009). In addition, Roche and Tirole (2003) require that (a) the two sides cannot coordinate and become a unified interest and (b) the amount charged by the platform on one side of customer cannot pass through to another side of customers.3 One-way indirect network effects. Armstrong (2006), Evans (2003b), Evans and Schmalensee (2007), and more recently Filistrucchi et al. (2013), have relaxed the above definition and basically consider sufficient that network effects exist for at least one group of customers. Armstrong is possibly the first to include both ‘membership’ (access) and ‘transaction’ (usage) types, whereas originally Roche and Tirole (2003) considered only transaction fees. According to Evans (2003b), the conditions for a two-sided platform are that (a) there are distinct groups of customers, (b) a member of one group benefits from having demand coordinated with one or more members of another group and (c) an intermediary can facilitate that coordination more efficiently than bilateral relationships between the members of the group. Evans and Schmalensee (2007) later add that for condition (b) above it is sufficient that one side of customers is attracted with the increasing size of the other. Filistrucchi et al. (2013) confirm and subscribe to this view and when discussing media platforms, while recognising that viewers generally do not like TV advertising, conclude that ‘it is not necessary for the existence of a twosided market that two indirect network effects be present. One suffices’ (Filistrucchi et al., 2013, p. 38). They characterise, for instance, the TV market as a 2SM with one positive and one negative indirect network effect.

3

The role of the pass-through is crucial in the analysis of a two-sided market (Weyl, 2010); when the two sides can coordinate to internalise the indirect network effect, this effect ceases to be an externality and one faces a situation of a firm selling complementary goods.

Platform Economics and the Sharing Economy    21 No need for indirect network effects. In their second contribution, Rochet and Tirole (2006, p. 657) considered what they termed the ‘cross-group externalities definition’ as ‘under-inclusive’ and proposed this definition where the key and only characterising element is apparently that the price structure is non-neutral: A market is two-sided (a two-sided platform exists) if the platform can affect the volume of transactions by charging more to one side of the market and reducing the price paid by the other side by an equal amount; in other words, the price structure matters, and platforms must design it so as to bring both sides on board (pp. 664–665). Therefore, it is only the price structure that matters, since indirect network externality, as a necessary condition, would make 2SMs under-inclusive. Yet, it must be observed that Evans and Schmalensee (2007) argue that for such a definition to be satisfied, the relationship between end users must be fraught with residual externalities, whereas according to Filistrucchi et al. (2013, note 7; 2014, p. 299), the definition proposed by Roche and Tirole is only apparently different from that provided by Evans and in practice is just broader. 2SM versus resellers versus vertical firms. In their approach to 2SMs and multisided platforms, Hagiu and Wright (2013; 2015a; 2015b, 2015c) contrast these to vertical integrated firms or resellers alongside the dimension of the amount of control exerted by the different players. These authors do not consider network effects as necessary and identify as the key features of two-sidedness and multisidedness because (i) they enable direct interactions between two or more distinct sides, and (ii) each side is affiliated with the market/platform. Direct interaction entails that sides maintain control over key terms of the interaction (pricing, bundling, delivery, marketing, quality of the goods or services offered, terms and conditions) as opposed to a situation where the intermediary takes control over such terms. Affiliation means that at each side the users make the investments needed to join the market/platform and interact with the other sides; such affiliation is needed to create cross-group network effects. Following this perspective, 2SMs are distinguished from resellers and VI firms, building on the theory of the firm and considering issues such as moral hazard and the trade-off between control and decentralisation (Fig. 2). Only when the two sides interact directly by way of their affiliation with the intermediary (market or platform), then we can talk of two- or multi-sided markets. When the intermediary buys from suppliers and sells to buyers, then direct interaction (and related control on, for instance, price) disappears and we face the condition of reseller. The interaction is direct to the extent that the sides retain residual control rights over the goods traded. This typically arises when the supplier retains ownership of the goods being traded: examples include eBay and shopping malls. In contrast, a pure reseller holds all residual control rights over the goods sold to the buyer. This typically arises when the re-seller takes ownership of the goods from the suppliers; old-fashioned retailers are typical examples. When an operator integrates and controls the provision of the services

22    Platform Economics

(A)

Sale of products or services

Suppliers “affilia on”

(B)

(C)

Suppliers

Suppliers

Sale of products Ver cal integra on (employment)

2SMs

Reseller

V.I. firm

Buyers “affilia on”

Sale of products

Sale of services

Buyers

Buyers

Fig. 2:  Two- and Multi-sidedness versus Resellers and V.I. Firms. Source: Adapted from Hagiu and Wright (2015c, p. 4). by professionals or workers, becoming directly responsible for them, then it is a VI firm (whereas in 2SMs the independent professional or worker retains responsibility for and residual control rights over the services). The fundamental trade-off in this strategic choice is between the coordination and control benefits that arise in a VI model (entailing also additional costs), and the benefits of lower costs from a two- or multi-sided strategy (at the cost of less control and of efforts needed to motivate professionals to adapt their decisions to the information that arises from the market/platform). In the VI mode there is a possibility of ‘moral hazard’ in that professionals may produce more efforts than needed; on the other hand, in the 2SMs mode there can be informationrelated moral hazard by online platforms that can extract insights from the aggregate data generated by the interactions between contractors and customers on their sites – insights that are not known to any individual contractor. It is worth noting that the way in which Hagiu and Wright (2015c) define direct interaction exclude some key sharing economy platforms and make them more similar to resellers or VI firms. In Uber and some digital labour platforms (i.e. especially Amazon Mechanical Turks), one side of the platform (i.e. the socalled ‘contractors’) has no control on several important dimensions; we come back to this at the end of this chapter and then again in Chapter 3 on digital labour platforms.

Classifications, Determinants of Size and Market Functioning Market-makers, Audience-makers and Demand Coordinators This classification has been formulated by Evans (2003b) and Evans and Schmalensee (2007), and was later used in several contributions that Evans (2009, 2011, chapters. 7–10) made focussing on online platforms. Considering the different transaction costs minimisation strategy, 2SMs can minimise transaction

Platform Economics and the Sharing Economy    23 costs by matchmaking (exchanges), building audiences (media) and minimising duplication costs (demand coordinators). They also note, however, that some new social media platforms can engage in all three functions. Market-makers (typically exchanges), by bringing together two distinct groups that are interested to transact, increases the likelihood of a match and reduces the time it takes to find an acceptable match (i.e. search costs). Each member of a group values the service more highly if there are more members of the other group. Audience-makers (off and online advertising-supported media) match advertisers to audiences; advertisers value the service more if there are more members of an audience who will react positively to their messages. Software platforms, operating systems and payment systems are defined residually as demand coordinators. They neither sell a transaction nor a ‘message’ (as market-makers and audience-makers) but coordinate demand and in so doing avoid duplication costs. In the remainder of this chapter, we focus our attention on the first category of market-makers (also referred to as matching markets) and on their digital version (although some scattered reference to physical 2SMs are made). With respect to audience-makers, there is another distinction between transaction and non-transaction types (which is worth mentioning as it could be applied by extension to some sharing platforms). According to Filistrucchi et al. (2013, 2014), whether a transaction is or is not performed is not a condition that can be used to qualify for two-sidedness: the media advertising business is a two-sided non-transaction market where there is no transaction between the two sides but there is a unilateral network effect as a sufficient condition. Leaving aside the different approaches seen earlier, matching markets in most cases and by most of the definitions illustrated earlier are 2SMs with clear bilateral and indirect network effects and price discrimination (Caillaud & Jullien, 2003a; Damiano & Li, 2008). A heterosexual dating club, for instance, is more attractive to both sides of the market as more women and men attend it, and usually charges a higher price from men. The logic is the same as in a totally different digital matching platform such as Airbnb, where having more guests increases the chances for hosts to rent their place, and having more hosts increases the chances for guests to find a nice place at a convenient price; the platform charges a higher transaction fee from guests (6–12%) as compared to hosts (3%).

Size Determinants Regardless of the types one is considering, from the literature one can find some ‘horizontal’ conditions that can affect platforms’ size and level of concentration (Table 1). Under clear network externalities the presence of indirect network effects promotes larger and fewer competing platforms. Only at some points and under specific circumstances, positive externalities from more participants may turn into negative externalities in the form of congestion. In many cases there is a fixed cost of providing a platform (i.e. investments over the years to solve the chicken and egg challenge and bring two sides on board), which means that economies of scale favour large platforms and concentration. In physical matchmaking platforms

24    Platform Economics Table 1:  Factors Affecting Platforms’ Size Determinants

Effects on Size/Concentration

Indirect network effects

+

Scale economies

+

Congestion



Differentiation



Multi-homing



Heterogeneity



Source: Our elaboration.

(i.e. dating club), there are evident physical limits to growth, beyond which the problem of congestion arises. Congestion increases search and transaction costs and, as we will see when dealing with new emerging P2P market-making platforms, it also affects digital platforms. Unless solved otherwise (increasing size of dating club, or improving search and matching algorithm in digital platforms), congestion limits growth and concentration. Platform vertical (i.e. upscale and downscale online market places) or horizontal (i.e. between Airbnb and Homeaway) integration reduces size and concentration and favour multi-homing, whereby on one or on both sides of the market, users choose to join and use several platforms. Heterogeneity of users and/or the object of exchange make matching more difficult and reduce the potential for scalability and concentration. So, digital platforms can generate strong network effects, reach scale and trigger competition policy questions, although as seen, scaling up to dominance (size) and industry concentration are constrained by congestion, heterogeneity and multi-homing.

Inherently Frictional Markets Digital matching platforms should aggregate and use information efficiently, while at the same time minimise the search and deliberation efforts to reduce transaction costs compared to other alternative channels and make it convenient for the users to transact on the platforms (Einav, Farronato, & Levin, 2015). In this optimisation problem, heterogeneity is a major challenge. Empirical research emphasises that P2P digital markets are inherently frictional (Fradkin, 2014; Horton, 2014). Recent work in this area has focussed on the microstructure of specific marketplaces, estimating search inefficiencies (Cullen & Farronato, 2015; Fradkin, 2014), heterogeneity in the matching process and problems of congestion (Horton, 2014), the consequences of search frictions and platform design for price competition (Dinerstein, Einav, Levin, & Sundaresan, 2014) and the differences between distinct types of pricing mechanisms (Einav, Farronato, Levin, & Sundaresan, 2013). With heterogeneity, matching buyers and sellers and defining personalised pricing are challenging and sharpen the trade-off between keeping transaction costs low and using information efficiently. Heterogeneity can be in terms of the type of users and/or the type of object exchanged: (a) TaskRabbit

Platform Economics and the Sharing Economy    25 has very high heterogeneity in both the users and the object exchanged; (b) Airbnb has higher heterogeneity in users and lower in the object exchanged; and (c) Uber has high heterogeneity in users only. Price can help deal with trade-off between transaction costs and efficient use of information to maximise match. Indeed, three empirical studies show that Airbnb (Fradkin, 2014), oDesk (Horton, 2014) and TaskRabbit (Cullen & Farronato, 2015) are characterised by high level of heterogeneity, frictions, a high percentage of non-matched potential (search friction) and congestion (i.e. a match fall through because of multiple requests at the same time).4 Apart from the pricing mechanisms, the use of data and search algorithms (i.e. data analytics) are crucial for digital platforms to increase their capacity to match the two sides of the market. In this respect, Einav et al. (2015) do not rule out that platforms can arrange search results and manipulate in a way that is more beneficial to them than to the users.

Centralisation Versus Decentralisation In addition to data analytics, pricing mechanisms can be leveraged to improve matching, and in this case the choice is between centralisation and decentralisation (Einav et al., 2015). A trend that marks a move from decentralisation to centralisation is the decline in the number of digital platforms that use auction mechanisms (Einav et al., 2013). TaskRabbit, for instance, abandoned the auction mechanism. Whereas earlier TaskRabbit accepted any possible task and let the taskers bid their price for performing them, now it accepts only standardised tasks that are offered at fixed prices; this move has been commented as TaskRabbit becoming the Uber for personal services.5 The continuum centralisation versus decentralisation can be illustrated by comparing Uber and Airbnb. This trade-off entails designing the platform as to centralise or decentralise choice. In the accommodation business, differences in preferences and in seller costs justify the decentralised design adopted by Airbnb. On the other hand, Uber needs to match in real time, especially in pick hours: the type of cars and the type of drivers are probably less important than getting a ride at a right time, which justifies its centralised design at least with respect

4

Fradkin (2014), for instance, reports that in Airbnb (a) potential guests typically view only a subset of potential matches in the market and more than 40% of listings remain vacant for some dates; (b) hosts reject proposals to transact by potential guests for 49% of the time, causing the potential guests to leave the market, although there are potentially good matches remaining; and (c) without search frictions (guests had all information and knew which hosts were willing to transact with them), there would be 102% more matches, and revenue per searcher would be $117 higher. In TaskRabbit, before the recent change of model (see next sub-paragraph), Cullen and Farronato (2015) found that the auction mechanisms are not very efficient as they do not vary much with market conditions, and suggested that a simpler mechanism may be preferable; this spot market clears due to suppliers’ elasticity: in periods when demand doubles, sellers work almost twice as hard, prices hardly increase and the probability of requested tasks being matched only slightly falls. Similar results are found by Horton (2014) for the oDesk market for professional services. 5 See, for instance, Newton (2014).

26    Platform Economics to the goal of maximising matches and revenues. The Uber surge-pricing algorithm let price per mile vary as supply and demand conditions change. In this case, the information collected and processed by the platform substituted for an auction mechanism. The flexibility and openness of Airbnb is reflected in a large variety of locations, prices charged and additional services provided by the hosts. It is one of the advantages and threats of Airbnb that it can make the service and experiences very diverse and rewarding, but at the same time, it makes the platform entirely dependent on the hosts of dimensions beyond control (apart from the system of ratings). Multi-homing is typical of Airbnb hosts that post their house in various platforms. On the contrary, Uber more rigidly imposes conditions on drivers; its offering is more standardised and controlled and the ride services platform has been trying to enlist providers in a more proactive manner. A final important difference is that multi-homing is more difficult (if not impossible) with Uber (and with Lyft). So, heterogeneity on both users and the object of transactions, frictions and the possibility for multi-homing seem to suggest at least ex ante that platforms such as Airbnb or TaskRabbit (in the original model) are less likely to scale up to dominance. This may be different for Uber, given its centralisation and standardisation and the little possibility for multi-homing.

On-demand Versus Scheduled Transactions The very specific characteristics of what is exchanged also distinguish digital match-making platforms in ways that are not entirely germane from a policy perspective. This aspect can be illustrated by using as source the written comment provided by the car-sharing platform RelayRides to the public consultation launched in the United States by the Federal Trade Commission in the run up to a workshop on the sharing economy held in June 2015.6 RelayRides makes a clear case for distinguishing its business model – defined as person-to-person carsharing platform connecting car owners to rent out their idle vehicles to travellers – from ride services business models such as Uber or Lyft; on these bases it calls policy makers not to take a one-size-fits-all approach to regulation (see Table 2). Table 2 compares Uber and RelayRides on the very specific characteristics of the object being transacted, and in our view the first one (on-demand vs scheduled transaction) has clear implications. As the exchange is scheduled and eventually entails a face-to-face meeting, it increases trust and add another layer of ‘protection’ on top of the reputational ratings system. It entails a relation between renter and owner (as in Airbnb) and not between drivers and passengers, which has clear implications in terms of liability.7 In general, 6

See  https://www.ftc.gov/system/files/documents/public_comments/2015/07/02031-96671.pdf. This also made possible for RelayRides to provide a $1 million liability policy for owners against injuries and third party property damage and protects owners against damage to their cars. Renters can choose to purchase different levels of insurance-protection packages at the time of rental – akin to traditional rental car practices. RelayRides has taken a proactive stance with respect to insurance and obtained commercial liability insurance policies for its community. 7

Platform Economics and the Sharing Economy    27 Table 2:  Car Sharing Versus Ride Services Car-sharing

Ride-services

Scheduled reservations

On-demand; short-lead

Driving for personal use

Driving for commercial use

Driver is third party (renter)

Driver is car owner

Longer, mostly downtime

Higher utilisation

Usually solitary

Usually multi-party

Owners and renters

Drivers and passengers

Source: Our elaboration.

and considering other digital match-making platforms (excluding those concerning labour and professional services such as TaskRabbit and/or Elance o’Desk), three possible dimensions of differentiation are as follows: (1) the extent to which the object of sharing is on demand or scheduled with some advance; (2) the frequency of occurrence; (3) the level of risk/safety. So, although the revenue stream is the same, one could certainly distinguish three different segments such as car-sharing (RelayRides), ride-sharing (BlaBlaCar) and ride services (Uber) that have different implications from the perspective of consumer protection. As the value of each transaction is higher in the first two segments compared to the latter one, there is less of a pressure to increase volumes and to forego consumer protection aspects. Less risk and betterdefined liability also make insurance policies easier to be defined. Although in a totally different sector, Airbnb is more similar to P2P car- and ride-sharing than to ride services.

Ratings and Related Governance Issues Exchange among strangers is one of the salient characteristics of many digitalmatching platforms; building trust to get both sides of a market on board has been a key challenge and driver of success for the biggest players. The ‘generalised trust’ (as distinct from particularised trust)8 that makes these new digital platforms thick is the combined working of users’ attitudes and how such attitudes are effectively leveraged by online reputational rating systems. The reliability of reputational rating is a policy and regulatory relevant topic, for some authors claim that these reduce information asymmetry and are a reliable form 8

Scholars of trust distinguish between generalised and particularise trust (Couch & Jones, 1997; Delhey, Newton, & Welzel, 2011; Freitag & Traunmüller, 2009; Putnam, 1993, 2000; Stolle, 2002; Yamagishi, Cook, & Watabe, 1998; Yamagishi & Yamagishi, 1994). Particularised trust concerns a close network of social proximity (i.e. family and friends). Generalised trust is a more abstract attitude towards other people and expectations about their behaviours, and can be defined as an attitude entailing reliance on the benevolence of human nature and to give most people the benefit of doubt.

28    Platform Economics of self-regulation, ensuring consumer protection and security that should not be altered by any form of regulatory intervention (Allen & Berg, 2014; Cohen & Sundararajan, 2015; Koopman, Mitchell, & Thierer, 2014, 2015; Sundararajan, 2014; Thierer et al., 2015). In practice, however, there are a number of reasons why such ratings may not be fully reliable. There is extensive literature on trust and reputation systems (Nosko & Tadelis, 2015; Pallais, 2013) which dates back to early works by Resnick and Zeckhauser (2002) and Bajari and Hortasu (2004). Online exchange faces intrinsically two sources of information asymmetry to be dealt with. The first source concerns the identity of counterparts (Ba, 2001; Pavlou & Gefen, 2004) and three forms of uncertainties: unidentified, anonymous and not possible to bind to a single person. The second source regards the quality of the object (goods or services) of exchange (Gefen, Benbasat, & Pavlou, 2008; Jøsang et al., 2007); the online consumer of a service or product most often needs to pay before receiving/ experiencing goods/services. Whereas identification and verification systems try and obviate to the first source, reputation ratings deal with the second one. Reputation ratings work on the assumption that reputation is a ‘value’ that can influence the capacity to exchange or sell a particular goods or services (Burnham, 2011). When the parties are total strangers to one another, reputation systems are collaborative filtering mechanisms helping the emergence of generalised trust (Corritore, Kracher, & Wiedenbeck, 2003). In a given community of online exchangers, reputational ratings are a sort of social control by which the members police themselves (Abdul-Rahman & Hailes, 2000; Ba, 2001). The second mechanism with the capability to increase trust in online marketplaces is the implementation of social networking features, or the leveraging of preexisting relationships (and by extension, existing pre-established trust) from the social graph of an individual. Such integration has two purposes in building online trust: confirming identity and establishing transitive trust (Abdul-Rahman & Hailes, 2000; Ba, 2001; Hogg & Adamic, 2004; Jøsang, Ismail, & Boyd, 2007; Kwan & Ramachandran, 2009; Swamynathan et al., 2008). The principle of trust transitivity refers to The idea that when Alice trusts Bob, and Bob trusts Claire, and Bob refers Claire to Alice, then Alice can derive a measure of trust in Claire based on Bob’s referral combined with her trust in Bob. (Jøsang et al., 2007, p. 624) There are, however, a number of shortcomings with reputational ratings that may undermine their reliability as a source of self-regulated consumer protection. There are basically two main potential biases: under-provision of ratings and strategic behaviour in provided ratings. Leaving an accurate rating is a public good and is likely to be under-provided (Avery, Resnick, & Zeckhauser, 1999; Miller, Resnick, & Zeckhauser, 2005). As a result, a given user may not always leave a rating, and the distribution of his/her evaluation may not accurately represent the outcomes of that agent’s previous transactions. It has been shown, for instance, with data from eBay that buyers and sellers with mediocre experiences

Platform Economics and the Sharing Economy    29 review less than 3% of time (Dellarocas & Wood, 2007). In two-sided reviews systems, users may provide more positive ratings than their true evaluation to avoid retaliation. When eBay had a two-sided review system, over 20% of negative buyer reviews were followed by negative seller reviews, interpreted by the authors as retaliation (Cabral & Hortacsu, 2010; Saeedi, Shen, & Sundaresan, 2015). On the other hand, it became evident through an experiment that a system in which reviews are hidden until both parties submit a review (‘simultaneous revealing’) reduces retaliation and makes markets more efficient (Bolton, Greiner, & Ockenfels, 2012). Fear of retaliation or intentional collusive behaviour with friends can lead reviewers not to reveal their experiences in the review. In various studies, it has been documented that some users who anonymously answered that they would not recommend their counterpart nonetheless submitted a public review giving a five-star rating. Furthermore, social communication can lead reviewers to omit negative comments due to two reasons. First, conversation can cause buyers and sellers to feel empathy towards each other (Andreoni & Rao, 2011). This may cause buyers to assume that any problem that occurs is inadvertent and not actually the fault of the seller. Second, social interaction may cause buyers to feel an obligation towards sellers because the sellers have offered services and were ‘nice’ (Malmendier & Schmidt, 2012). This obligation can lead buyers to omit negative feedback because it would hurt the seller or it would be awkward. Whatever the sources, these biases may reduce market efficiency and, for example, may cause users to engage in suboptimal transactions (Horton & Golden, 2015; Nosko & Tadelis, 2015). Some empirical contributions focussing on reputational ratings with respect to the new P2P digital-matching platforms corroborate above considerations (Cullen & Farronato, 2015; Fradkin et al., 2015; Horton & Golden, 2015; Lauterbach, Truong, Shah, & Adamic, 2009; Overgoor, Wulczyn, & Potts, 2012; Zervas, Proserpio, & Byers, 2015). The first two studies focussed on CouchSurfing, and using big data scraped from the web conclude that there is a bias towards positive reviews and that there can be collusive reciprocity among individuals belonging to the same network (Lauterbach et al., 2009; Overgoor et al., 2012). A comparison of the distribution of reviews for the same property on TripAdvisor and Airbnb shows that ratings on TripAdvisor are lower than those on Airbnb by an average of at least 0.7 stars (Zervas et al., 2015). More generally, the rate of five-star reviews is 31% on TripAdvisor and 44% on Expedia (Mayzlin, Dover, & Chevalier, 2014) compared to 75% on Airbnb. This difference in ratings could be interpreted as showing that two-sided review systems induce bias in ratings. A recent study involving researchers affiliated to Airbnb document through field experiments conducted on Airbnb itself that there are some biases, but when such biases are removed through experimental manipulation, the five-star ratings on Airbnb remain substantially higher than 44% (Fradkin et al., 2015); this would imply that these are a reliable measure of quality to inform other consumers. The study of another platform (Elance oDesk) documents through a laboratory experiment that reputational ratings are fairly inflated (Horton & Golden, 2015). The evidence is, thus, at best inconclusive and mixed on whether or not reputational ratings are a sufficient and reliable measure of quality and consumer protection.

30    Platform Economics The limitation of ratings as a form of self-regulated consumer protection brings us into the debate on the governance and regulation of digital matching platforms. This debate is polarised between those radically against any intervention (Allen & Berg, 2014; Cohen & Sundararajan, 2015; Koopman et al., 2014, 2015; Sundararajan, 2014; Thierer et al., 2015) and those that are in favour of some forms of regulation (Cannon & Chung, 2015; Gobble, 2015; Malhotra & Van Alstyne, 2014; McLean, 2015; Ranchordas, 2015; Rauch & Schleicher, 2015; Sunil & Noah, 2015), with some more specialised legal approaches (Barry & Caron, 2014; Cohen & Zehngebot, 2014; Daus & Russo, 2015; Miller, 2014, 2015; Oei & Ring, 2015) proposing in some cases very strict interventions such as on taxes (Oei & Ring, 2015) or on transportation services (Daus & Russo, 2015). The first position is utterly libertarian and counter-argues the market failure rationale for regulating with the argument of regulatory capture and regulatory failure so that the self-regulating nature of the market is a lesser evil (Allen & Berg, 2014; Koopman et al., 2014; Thierer et al., 2015). From this standpoint, excessive legislation and regulation could absorb and neutralise consumers and efficiency gains produced by technological innovation. The Internet and the rapid growth of online platforms alleviate the need for top-down regulation, as these recent innovations are likely doing a much better job of serving consumer needs. It is actually argued that reputational feedback mechanisms (the ratings) solve the classical information asymmetry that goes under the well-known heading of ‘lemons problem’ (Thierer et al., 2015). These authors call for a new bottom-up self-regulated approach where: (a) various forms of licencing should be reduced to allow private certification schemes and reputation mechanisms to evolve; (b) regulations making it difficult for start-ups to compete for labour (contractors should not be turned into employees) should be avoided; and (c) regulation should remain general and not industry-specific. More nuanced and less radical approaches call for some form of regulation attempting a compromise to ensure consumers’ protection and safety without stifling innovation (Barry & Caron, 2014; Miller, 2014, 2015; Ranchordas, 2015; Rauch & Schleicher, 2015; Sunil & Noah, 2015). Largely, they propose ‘smarter regulation’ envisaging a number of possible solutions such as, for instance, the use of information-based regulation (metrics and performance); this, however, would require transparent and updated data on the functioning of these digital intermediaries, which the platforms seem unwilling to provide. A very balanced position expressed not in a regulatory essay but rather at the end of an economic analysis of P2P market is that of Einav et al. (2015). First of all, the authors recognise that the welfare and labour effects of these platforms remain an empirical question, which can be interpreted as saying that ex ante positions in any direction (in favour or against regulating) are not yet empirically grounded. They also clearly recognise that ratings can be biased and inflated and that it is possible that platforms present the results of search in a way that is more convenient to them than to the users. On the other hand, they point out that imposing licencing and certification on the platform may protect incumbent without really protecting consumers; although such requirements can be seen as remedies to market failures, their implementation takes the form of lengthy

Platform Economics and the Sharing Economy    31 processes after which little monitoring is performed. In this respect, they seem to favour small interventions allowing traditional industries and new platforms to compete on an equal footing. With respect to the utilisation of data by the platforms, they observe that several questions emerge, such as follows: Can consumer limit platform use of data? Can platform share/sell ratings and purchase history? What about potential gender and race discrimination in ratings leading to lower opportunity for certain groups? In this respect, it is worth noting that empirical studies have shown that online platforms and their functioning may produce some form of discrimination. A statistical analysis of a dataset constructed from Airbnb (combining pictures of all New York City landlords on Airbnb with their rental prices and information about quality of rentals) finds what can be seen as indirect evidence of racial discrimination (Edelman & Luca, 2014). The main finding is that, controlling for other relevant covariates, non-Black hosts charge approximately 12% more than Black hosts for equivalent rentals. These effects are robust when controlling for all information visible in the Airbnb marketplace. Another empirical study based on Airbnb data (using both platform data and doing a controlled experiment) found that the more trustworthy the host is perceived to be from her photo, the higher the price of the listing and the probability of its being chosen and that a host’s reputational ratings have no effect on listing price or likelihood of consumer booking, even when such ratings are manipulated experimentally (Ert, Fleischer, & Magen, 2016). The impact of just a photo, although not full-fledged discrimination, may at least suggest that quality based on ratings is not such a bulletproof instrument of consumer protection based only on rational criteria.

Back to ‘Sharing Platforms’ Some of the digital platforms that currently present themselves as part of the ‘sharing economy’ are evidently 2SMs by all standards: network effects, price non-neutrality, direct interaction (retaining control over key terms of exchange) and platform affiliation. Obviously, not all sharing platforms qualify as 2SMs. An early champion of the sharing economy such as Zipcar is a pure reseller. It owns fleets of cars that it rents to customers, albeit digitally and following an innovative business model, unlike platforms, which only facilitate transactions between car owners and car users. Nonetheless, a large majority, especially among the most successful ones, are 2SMs. If we take Airbnb as an example: (a) more hosts (suppliers) will attract more guests (consumers) and vice versa; (b) hosts and guests are charged different transaction fees (3% from hosts, and between 6% and 12% from guests); (c) hosts retain full control over choosing when their room or apartment is available over the price and other aspects (the platform only makes recommendations); and (d) both hosts and guests make necessary ‘investments’ to become affiliated with the platform (although multi-homing is widely practised on both sides). Starting from this observation, rather than from the alleged ‘sharing’ and communitarian elements, would have avoided much of the current rhetorical obfuscation with a number of relevant implications. First, commercial sharing platforms’

32    Platform Economics revenues depend entirely on volume (fee on transactions, and subscription-based models are rarer), which may lead to maximisation of transactions even to the risk of creating negative externalities (i.e. not sufficient security and safety checks, over-crowing of neighbourhood with tourists, etc.). Second, not for profit, ‘true sharing’ platforms may be analysed as two-sided exchanges, even if no monetary transactions are involved. They would still need to bring somehow on board two sides, and their scaling up would depend on achieving direct and indirect network effects. Using such perspectives, one would easily spot that many of these `true sharing’ platforms (except cases as CouchSurfing) remain small and do not scale up, which could support an investigation into what makes the difference in this respect with commercial sharing giants. Third, considering the empirical evidence, frictions would easily moderate the claims about the market efficiency that are part of the rhetoric on the promises of sharing platforms. Fourth, evidence on the limits of reputational ratings can easily be used to challenge the claims about smooth self-regulating sharing platforms. There are two additional considerations: one on competition policy and another on the issue of control and labour relations; these are devoted relatively more space below. There are possible implications of ‘sharing’ platforms for competition (King, 2015; Lougher & Kalmanowicz, 2016). Competition issues first emerged in Europe as a result of complaints filed by Uber and Airbnb against restrictions imposed in certain member States. This has been confirmed in the hearing that both companies had with the UK House of Lords (2016, pp. 25–50). According to King (2015), there are three potential concerns: (1) anti-trust implications when platforms activate network effects leading to dominance; (2) lock-in of third parties on one side of the transaction; and (3) power to reference rivals (i.e. leading to collusion or alternatively to discriminatory behaviour). King (2015, p. 731) explains that Uber has also faced pricing complaints for its ‘surge pricing’, which was voluntarily limited in New York due to the potential of such pricing to breach price-gouging laws. In the review by Lougher and Kalmanowicz (2016), it is not ruled out that ‘sharing’ intermediation markets can become concentrated and possibly dominated by a single market player. The activities of powerful platforms, for which data usage is a key, are likely to be scrutinised in merger control proceedings, and in the long term potentially also in the area of market abuse. They cite statements by representatives of French and Germany competition authorities to substantiate the claim that market power for such platforms comes from the capacity to collect a large amount of personal data and use it commercially (Lougher & Kalmanowicz, 2016, pp. 96–97). Finally, after noting that the regulation of ‘sharing’ platforms is hotly debated, they report that although some member states have called for specific regulatory framework, however, the European Competition Commissioner Margrethe Vestager made clear in several statements that such platforms are too diverse to be monitored through a single regulatory framework and that it is preferable to apply existing antitrust rules case-by-case (Lougher & Kalmanowicz, 2016, p. 102). From the evidence that we reviewed on the characteristics and functioning of the largest platforms, it seems that market dominance is out of reach for most of them due to heterogeneity and matching frictions, but is not so unlikely for Uber. On the other hand,

Platform Economics and the Sharing Economy    33 improvements in the matching algorithms, together with pricing strategies and use of personal data without any regulatory checks, may change the situation and make market dominance more likely for a few other platforms. Finally, the issue of control emerges from the earlier considered perspectives presented in the works of Hagiu and Wright (2013, 2015a, 2015b, 2015c), on which it is worth coming back to anticipate some elements that will be further discussed later, especially when dealing with digital labour platforms in Chapter 3. As explained, for these authors, it is the direct interaction between sides that sets 2SMs apart from both resellers and fully VI firms. Direct interaction presupposes that platform users retain control on some of the key terms of this interaction such as ‘pricing, bundling, delivery, marketing, quality of the goods or service offered, terms and conditions’ (Hagiu & Wright, 2015c). Using this dimension, at one extreme there are cases such as Avis and Zipcar, for they own fleets of cars that they rent out to consumers. On the opposite extreme, newer services, such as Getaround, Lyft and RelayRides, simply facilitate transactions between car owners and car users. None of these services owns a fleet, and the car owners set their own rates except in the case of Uber, which imposes some standard rates, and for this reason is placed in the picture aside from the pure 2SMs types. This issue of control is not germane from a policy perspective as it points to the debate on whether providers performing work tasks should be considered independent contractors or de facto employees. Fig. 3 conveys the trade-off between control and cost comparing pure marketplaces with pure integrated firms and suggest the strategy that some sharing platforms have been trying to enact that are sources of conflicts and policy concerns. Control, which is important in all two-sided strategies, is even more critical when dealing with labour. There are various reasons for this: typical matching frictions, the heterogeneity of tasks/contractors/employers, prominence of on-demand and time-sensitivity (i.e. Uber), and problems of coordination of

Cost structure

Integrated firm (employees)

Pure marketplace

(independentcontractors)

Some sharing plaorms? Control

Fig. 3:  The Control and Cost Trade-off. Source: Our elaboration.

34    Platform Economics multiple contractors. Obviously, control is maximised in vertically integrated firms to ensure consistency, speed, timely delivery, coordination and scale. However, control has a cost: employees have higher costs than independent contractors. On the other hand, lower costs mean less control, although some of these platforms seem to be striving to minimise costs and maximise control, almost to the level typical of a vertically integrated firm. The attempt of achieving both lower costs and control through algorithm-driven management raised important conflicts and regulatory issues that will be addressed in Chapter 3.

Chapter 2

Rhetoric, Reality, Impacts and Regulation in Labour Intermediation Platforms Introduction This chapter focuses on the ‘sharing economy’ in general, including also in this umbrella concept those platforms intermediating labour that are discussed in depth in Chapter 3, by introducing and discussing three important elements. The first is the conceptual ambiguity of the expression performing with respect to the second element, which is the use of rhetorical discourses as a framing weapon by the conflicting sides; these are contrasted against the available empirical evidence. Third, from the above elements, the relevant regulatory and policy themes are extracted and anticipated by way of considering selectively some important legal disputes. As the dust of the hype settles, policy makers are starting to see through the obfuscation produced by the terminological ambiguity of expressions such as sharing economy (and all others alternatives: collaborative consumption, collaborative economy, gig economy etc.) and calling for more neutral terms, exactly as we have done in both Introduction and Chapter 1. A report by the European Parliament underscored the value loadedness connotation of the expression ‘sharing economy’ and backed the decision by European Parliament’s Committee on Employment and Social Affairs to use ‘platform economy’ and ‘platform work’ as the most neutral expressions for the activities placed under the ambivalent expressions currently in use (European Parliament, 2017b, p. 21). Similarly, a report published by the European Agency for Safety and Health at Work (Garben, 2017, p. 13) uses the term ‘online (labour) platform work’ instead of sharing crowdwork platforms. The same report, however, shows how in many countries the expression ‘sharing economy’ is still dominant in the public debate, although some stakeholders’ group, such as, for instance the Confederation of Danish Trade Unions,1 have encouraged the use of ‘platform economy’ as a more neutral term. Inconsistent and incoherent use of different terminology and lack of clear-cut definitions and conceptualisation have caused both lack of reliable and consistent 1

From Rasmussen and Kongshøj Madsen (2016). Reported in Garben (2017, pp. 42–43).

Platform Economics: Rhetoric and Reality in the “Sharing Economy” Digital Activism and Society, 35–71 Copyright © 2019 by Emerald Publishing Limited All rights of reproduction in any form reserved doi:10.1108/978-1-78743-809-520181002

36    Platform Economics in the size of the phenomenon (how much turn over, how many people involved) and elaboration of policy approaches and regulations targeting a well-identified set of players. Although in this book we also take self-defining practices at face value as object of empirical analysis and have anticipated that we prefer to define commercial sharing platform as two-side market (or in cases where they exert close control as de facto vertically integrated firms), nonetheless we reconstruct and deconstruct this conceptual ambiguity for two reasons. First, because tracing the origin of terminological use is closely linked to discussing rhetorical framing and lobbying. Second, because clarifying certain distinctions can at least set the grounds for a more informed debate of policy and regulation. Self-defined ‘sharing platforms’ basically match different groups of users and providers and enable the increase in scale and speed for traditional transactions such as selling, renting, lending, labour trade, etc. Yet, they do it in different ways and following different business models that have very different policy and regulatory implications. It is pointless to discuss, for instance, BlaBlaCar and Uber, or TaskRabbit and Time Banks,2 as part of the same domain. Grouping highly profitable companies such as Airbnb and Uber alongside voluntary gift-giving exchanges such as Freecycle or CouchSurfing contributed to fuel conflicting rhetorics and controversies. It comes as no surprise that without a clearer and consensual definition, no reliable measurement of the phenomenon’s dimensional relevance exists. As a corollary to this conceptual analysis, we will also briefly review some of the available and inconsistent quantitative estimates of the phenomenon. On the basis of the reconstruction of deconstruction of terminological and conceptual ambivalence, we then move to tracing the initial emergence of the ‘sharing’ discourse and movement and how from this initial stage different rhetoric emerged. After the development of ‘sharing’ platforms took a more ‘commercial turn’ but retained its ‘sharing’ rhetoric, disenchantment fuelled growing criticism, which was subsequently exacerbated by the mobilisation of more tangible interests and concerns (i.e. those of the disrupted traditional interest) and by court cases and the first timid policy and regulatory interventions. We have identified rhetorics and discourses into ‘social utopianism’, ‘business and economics-driven optimism’ and ‘social pessimism’. From these three strands, we select and discuss the most controversial themes, which we contrast against available empirical evidence. In between the analysis of rhetorics and of empirical evidence, we also show how rhetoric and evidence have been used in what we call ‘lobbying as framing’. Additionally, in Chapter 4, we conduct an ideological production analysis peculiar to sharing economy players (neoliberal, commons and platform cooperativism discourses), which emerged from fieldwork in Barcelona, Paris and Berlin. Finally, it descends smoothly from the previous two elements: the final part of this chapter anticipates and reviews briefly the interests at stake and the 2

Time banks are initiatives that emerged in the 1980s and involved community-based t­rading of services on the basis of the time spent and following the principle that every member’s time is valued equally (Cahn & Gray, 2015; Cahn & Rowe, 1992; Collom, Lasker, & Kyriacou, 2012).

Labour Intermediation Platforms    37 corresponding policy and regulatory issues, which are reconsidered in the conclusive Chapter 5. In view of the ongoing debates and disputes and the potential of these platforms for disruptive economic and social innovation, it should come as no surprise that there are several interests at stake: (a) users as consumers (who supposedly receive large benefits from cheaper and more convenient choices because of more competition. However, they may also face risks due to lack of consumer protection and clear liability rules); (b) users as providers (i.e. the alleged ‘micro-entrepreneurs’ who drive the cars, let their homes or perform errands using different platforms. This is the most diverse and controversial group from a policy perspective since economic opportunities may be offset by concerns about erosion of workers’ rights); (c) the platforms (the owners of the ‘sharing economy’ platforms who have much to gain or lose, depending on future regulatory decisions); (d) established operators (i.e. operators in potentially disrupted industries such as taxi drivers, who stand to lose the most if the ‘sharing economy’ remains unregulated) and (e) the general public interest that may be benefitted from positive externalities or bear the costs of negative externalities (the ‘sharing economy’ can have positive or negative spillovers also for the economy and society as a whole, as it does, for instance, with the positive externalities of innovation, or with security risks or the alleged erosion of the broadly defined labour contract and the tax base).

Trajectory and Conceptual Issues Conventionally, the emergence of the sharing economy has been dated to 2008, which is the founding year of Airbnb and marks the pick of the financial and economic crisis (Schor & Fitzmaurice, 2015). Although short time has passed since 2008 and the lack of data put any attempt to identify the causal factors leading to the development of sharing economy and to forecast its future beyond the domain of science, some have attempted to identify drivers and the future trends. Although the distinction between rhetoric and analysis in such sources is a fine line, we think worth nonetheless to briefly review these alleged drivers. First of all, obviously there is technology in general, and in particular the kind of technology that came of age in the last 20 years of industry expertise in designing market places, that is to say, search and matching algorithms together with reputational ratings (Edelman & Geradin, 2016; Horton & Zeckhauser, 2016b). The technologies have abated almost to zero, the so-called Bring-to-Market (BTM) costs for individuals, and harnessed the trust mechanisms supporting transactions and collaboration among strangers. Beyond this clear and undisputable driver, the other is object of reasonable narratives begging, however, empirical evidence. The economic crisis that started in 2007 is mentioned among one of the drivers.3 The explosion of ­urbanisation 3

In a Delphi study involving 25 experts (Barnes & Mattsson, 2015), the most cited driver for the sharing economy was the need to save in the context of the crisis (followed by technology and socio-cultural changes). Some have argued the fact that the crisis that started in

38    Platform Economics and its burden on logistics and mobility is presented as another driver for sharing.4 Socio-cultural trends are also seen as fuelling and being fuelled by new technological possibilities that enable the alleged new consumer preferences for access over ownership5 and the engineering of generalised trust of strangers.6 Finally, environmental concerns and trade imbalances push toward the recirculation of goods, which sharing platforms enable.7 In brief, this narrative not immune from rhetoric and framing goes as follows: New technological possibilities and the economic crisis that started in 2008 have triggered participation of individuals in platforms as consumers and providers of rental services and/or

2007 and continued for the next eight years is among the key drivers for sharing economy boom; see, for instance, an academic (Schor, 2014) and a media (Roose, 2014b) discussion of this kind of explanation. 4 Cohen & Muñoz (2015) stress the combination of increased urbanisation and the need for new forms of sustainable production and consumption. Urbanisation is mentioned often as a driver. In 2008, for the first time in history, more people were living in cities than in rural areas, and by 2030, 5 billion people will live in urban areas (Dobbs et al., 2011, 2012; UNDESA, 2014). According to a World Economic Forum (WEF, 2013, 2014), urbanisation puts great pressure on mobility and logistics and creates a further push to increase access to shareable assets and to go towards circular and shared modes of both consumption and production. 5 Various consumption theorists see the irreversible advent of the switching from ownership to access as consumers are becoming more comfortable with this practice and with sourcing trust through peer-review systems (Baumeister & Wangenheim, 2014). This trend is allegedly reinforced by various surveys cited in policy and industry reports (Barbezieux & Herody, 2016; European Economic and Social Committee (EESC), 2014; Observatorio Cetelem, 2013; PIPAME, 2015). In addition, according to the more optimist and utopian analysts, there is an increasing desire for community (Gansky, 2010) and for richer small world experiences (Owyang, 2013, p. 5). 6 Platforms technology is seen as having impact on social capital and trust. For instance, whereas digital natives seem to trust people less (Pew, 2014), when they have used social networking sites, they become three times more likely to think that most people are trustworthy, and this also applies to other population groups (Hampton, Goulet, Rainie, & Purcell, 2011). It is further argued that connected consumers are much more likely to trust strangers online and, as they will become the overwhelming majority of the population, this ensures that the sharing economy would further develop and consolidate (Vaughan & Hawksworth, 2014). 7 The previous drivers are thought to interact with other behavioural and economic aspects. In France, for instance, it has been estimated that potentially shareable goods account for about 25% of expenditure and for about one-third of household waste (Demailly & Novel, 2014); alternatively, it has been calculated that on average each French family holds 70 unused objects and this makes a potential recirculation market worth €12 billion (PIPAME, 2015, p. 27). Imbalances among different parts of global economy have created and will continue to create accumulation of cheap imports in wealthy nations, which is seen as one of the drivers for the uptake of platforms enabling the recirculation of goods (Schor, 2014; Shor & Fitzmaurice, 2015). Developments at socio-cultural and socio-economic levels are then seen in clear relation with the environmental pressures that allegedly will make the transition from consumption to access irreversible (Vaughan & Hawksworth, 2014; World Economic Forum (WEF), 2014).

Labour Intermediation Platforms    39 sellers of goods (imbalance between rich and poor countries with inflows of cheap goods to the former have produced accumulation of unused goods in the former). The same process occurred for the offer and use of on-demand labour traded in platforms. Such socio-economic drivers are co-evolving alongside sociocultural changes with increasing preference for access-based consumption (as opposed to ownership) and to some extent for more flexible forms of employment. The concentration of billions of individuals in large cities puts mobility, logistics and space systems under strain, and here again platforms enabling new forms of mobility, accommodation, and delivery are playing and will play a role. Finally, environmental pressures and sustainability objectives have encouraged individuals to use resources more parsimoniously, which add a further driver for more efficient forms of consumption. As stated, there is no conclusive empirical evidence supporting this view of what caused the emergence of sharing economy and the underlying implicit or explicit narrative about its future growth. This narrative, however, is instrumental in reconstructing and deconstructing terminological use and later in the discussion of rhetorical discourses and in the contrasting of these with the available empirical evidence. The first conceptual framing, or at least the most popular original one, was that of ‘collaborative consumption’ by Botsman and Roger (2010) and emphasised the leveraging of idle assets. According to these authors, the concept is defined as including ‘bartering, lending, renting, gifting and swapping’ and is further divided into the following three categories: ‘product service systems’ (access to products or services without the need for owning the underlying assets), ‘redistribution markets’ (i.e. re-allocation of goods) and ‘collaborative lifestyles’ (i.e. exchange of intangible assets). Airbnb and Uber started to monetise idle capacity (space and cars), other early platforms promoted non-monetised loaning (i.e. Yerdle, Landshare, NeighborGoods and Share Some Sugar), still others launched second-hand markets, both monetised (i.e. swap style) and for free (i.e. Freecycle), while there were also several community-based platforms based on NFP-sharing and cooperation (Schor & Attwood-Charles, 2017). Criticising this definition (Belk, 2014b) and distinguishing between ‘true’ and ‘pseudo-sharing’, Belk (2014a) defines (a) collaborative consumption as ‘people coordinating the acquisition and distribution of a resource for a fee or other compensation’; and (b) ‘true sharing’ as entailing temporary access rather than ownership, no fees or compensation, and use of digital platforms. Belk (2014a) clearly makes the point that the majority of commercial platforms are improperly included in the ‘sharing economy’. At some point, in fact, the expression ‘sharing economy’ emerged, substituted collaborative consumption and became institutionalised (Schor, 2014).8 Sharing economy is the most used and widespread expression, although there 8

A group of platform founders, consultants and non-profits formed a ‘sharing’ listserv. In France, the OuiShare group was founded that became global and started organised sharing festivals (see brief history at https://handbook.ouishare.net/a-brief-history-of-ouishare). In Chapter 4, we include interviews with the five participants of the OuiShare festival in Paris, April 2016.

40    Platform Economics are still many others that more or less are used to refer to the same kind of platforms (collaborative economy, access-based consumption, gig economy, peer economy, on-demand economy. All these expressions are used as ‘floating signifiers’ for a diverse range of activities (Nadeem, 2015, p. 13)). The European Commission (2016b) in its communication continued to use a value-loaded expression such as ‘collaborative economy’. In view of the need for more neutral policy and regulatory terminology underscored at the beginning of this chapter, it is telling of the rhetorical climate the fact that still in 2016 an important policy and regulatory player such as the Commission used a value and rhetorically loaded expression. This and other instances of usage of such expressions by regulatory bodies is indicative of the enduring effects that the initial phase of ‘naming’ and ‘framing’ has had for both later emergence of rhetorical disputes and analytical confusion and ambiguity. In practice, ‘the definition of the sharing economy became a matter of self-selection by participating entities’ (Schor & Attwood-Charles, 2017, p. 3). So, the expression rather than being connotative became performative (Frenken & Schor, 2017; Richardson, 2015). Large commercial platforms were benefitted from association with the socially positive values of community and sharing, whereas for not-for-profit (NFP) platforms it was attractive to meet at the same events with powerful and richly endowed players (Schor & Attwood-Charles, 2017, p. 3). After the terminological fair started and platforms self-placed themselves into one or more labels, there have been various attempts to come up with some more coherent definitions and/or typology. However, this self-define practice makes it hard even for scholars to formulate ‘externally rigorous’ definitions. In some cases, authors pragmatically accept this semantic confusion that characterises practice. They are satisfied for the purposes of their inquiries to consider the ‘sharing economy’ as comprising ‘peer-to-peer Internet platforms (including Airbnb, Uber, TaskRabbit, Just Park, …), which empower individuals to monetise their underutilised assets, time and skills’ (Martin, 2016, p. 153). Industry analyst Jeremiah Owyang (2015) define the ‘collaborative economy’ as ‘an economic model where technologies enable people to get what they need from each other – rather than from centralised institutions’. Another expression used is ‘access-based consumption’, defined as ‘transactions that can be market mediated but where no transfer of ownership takes place and differ from both ownership and sharing’ (Bardhi & Eckhardt, 2012). A similar approach is used to define the ‘sharing economy’ as ‘consumers (or firms) granting each other temporary access to their under-utilised physical assets (‘idle capacity’), possibly for money’ (Frenken, Meelen, Arets, & van de Glind, 2015; Meelen & Frenken, 2015). On the basis of qualitative field work and the review of 254 platforms, collaborative consumption has also been defined as ‘a peer-to-peerbased activity of obtaining, giving, or sharing the access to goods and services, coordinated through community-based online services’ (Hamari, Sjöklint, & Ukkonen, 2016). Schor and associates have defined the ‘sharing economy’ as digitally ‘connect consumption’ to convey the importance of the digitally mediated social component and make a distinction from earlier forms of sharing and

Labour Intermediation Platforms    41 collaboration (Dubois, Schor, & Carfagna, 2014; Schor, 2014, 2015; Schor & Fitzmaurice, 2015; Schor, Fitzmaurice, Carfagna, & Will-Attwood, 2014). Starting from this premise, they have provided slightly different formulation briefly reported here. Schor (2014, p. 2) presents the following illustration in terms of broad categories: ‘Sharing economy’ activities fall into four broad categories: recirculation of goods, increased utilisation of durable assets, exchange of services, and sharing of productive assets’. It is worth pointing out the evident contradiction between defining it as ‘connected consumption’ and including ‘sharing of productive assets’. In Schor and Fitzmaurice (2015, p. 415), the four categories are ‘re-circulation of goods, exchange of services, optimising use of assets and building social connections’. Starting from the latter categorisation, Schor’s (2015, p. 14) last definition of what she calls the ‘new sharing economy’ is as follows: ‘Economic activity that is peer-to-peer, or person-to-person, facilitated by digital platforms’. These broad-based categories are a hybrid mix of factor markets (goods and labour) with specific sectors (accommodation and transportation). More consistently, one could identify three broad categories and match them to a traditional economic classification as follows: (a) recirculation of goods (second-hand and surplus goods market); (b) increased asset utilisation (production factors markets) and (c) service and labour exchanges (labour market). These attempts at defining and conceptualising the ‘sharing economy’ do not fully consider all the elements that we deem important for our analysis of both rhetoric and policy and regulatory implications. First, one of the key rhetorical and ideological discourses about ‘unlocking the potential of idle assets’ is used as a key definitional element without empirically questioning the differences in the kind of asset used and to what extent this asset is actually underutilised. The next chapter on digital labour markets explains that (1) leveraging a property or just labour makes a difference in terms of distributional and employment effects and (2) in some cases, it is not a matter of using free time for ‘pin money’ but a way of making necessary income. Second, differences in the interaction modality (peer-to-peer (P2P), business-to-consumer (B2C) etc.) are not fully considered. Digital platforms that define themselves as part of the ‘sharing economy’ include cases of B2C transactions: one of the early ‘sharing champion’ such as Zipcar is not a two-sided market (2SM) but a reseller and despite the self-appointed label it is an activity that is already fully regulated as opposed to other P2Ps or peersto-businesses (P2Bs) platforms operating to some extent, yet not fully regulated. Third, the distinction between commercial and NFP platforms, an aspect that is related to current rhetorical battles, has to be fully factored in. Fourth, platforms diverge in terms of dimensional relevance (from a few hundred users to millions of users). Finally, and most importantly, there are differences that cut across these broad-based categorisations in terms of regulatory and policy implications. Platforms placed in the same broad category (for instance, Uber and BlaBlaCar) differ widely in terms of their current regulatory implications (e.g. market access and licencing, liability and insurance, consumer protection and labour laws) and their potential to disrupt incumbent industries. It is worth pointing out that this

42    Platform Economics is also the perspective expressed (orally or in writing) in public consultations by representatives of, for instance, Relay Rides9 and Airbnb and Uber.10 Recently, Schor and Attwood-Charles (2017) have critically reconsidered all previous definition with some scepticism, including the use of the P2P dimension, a key distinguishing features, for they observed that many large platforms exert tight control that reduce the autonomous interaction among peers. Commenting on the economists’ approach to consider platforms as 2SM and multi-sided markets, these authors conclude that this practice is not satisfactory, as sharing sector is only one type of multi-sided market. Yet, it is our view that they missed the main point that we made in Introduction and Chapter 1, and basically ended up accepting at face value that commercial sharing platforms have something intrinsically distinctive from other platforms that are considered as 2SM- or multi-sided markets. One thing is, in fact, to study sharing discourses and practices in their own right and accepting self-definition (as we do below), and another thing is to analytically conclude that self-defined commercial sharing platforms are radically different from 2SM. The only difference is, however, the object of intermediation that could be the basis of sectorial categorisation; the dynamics and business model of commercial sharing platforms are not at all dissimilar from other 2SM, except possibly in the degree of control that raised about platforms, such Uber, being more similar to a vertically integrated firm than to a 2SM (see Chapter 1, and later again in the present chapter about this issue). In order to take self-defining practice at face value and at the same time distinguish a few broad types, we adopt the following definition formulated in Codagnone, Abadie, and Biagi (2016b, p. 22): The expression sharing economy is commonly used to indicate a wide range of digital commercial or non-profit platforms facilitating exchanges amongst a variety of players through a variety of interaction modalities (P2P, P2B, B2P, B2B, G2G) that all broadly enable consumption or productive activities leveraging capital assets (money, real estate property, equipment, cars, etc.) goods, skills or just time. 9

In a written comment sent to a public consultation launched in the United States by the Federal Trade Commission Relay Rides makes a clear case for distinguishing its business model – defined as a person-to-person car-sharing platform, connecting car owners who rent out their idle vehicles to travellers – from ride services business models such as Uber or Lyft. On this basis, the platform asked policy makers not to take a ‘one size fits all’ approach to regulation. The comment can be retrieved at https://www.ftc.gov/system/files/ documents/public_comments/2015/07/02031-96671.pdf. 10 In a written evidence, Airbnb presented to the UK House of Lords (2016, p. 25) inquiry into the Commission DSM strategy; it explicitly stated that given the diversity of business models ‘an effective single regulatory framework suited for such varied models would be impossible to conceive … there are significant differences that require bespoke approaches’. This aspect was reiterated in public through oral auditions by Patrick Robinson, Head of Public Policy Europe and Canada, Airbnb, Mark McGann, Head of Public Policy EMEA, Uber (House of Lords, 2016, pp. 37–50).

Labour Intermediation Platforms    43 P2P

P2B P2Bmoney lending

NFP

FP

Trues haring plaorms not a ma­er of regulatory concern

Commercial B2C already regulated (mostly not twosided markets)

Low skilled labou intensive online service (rouƒne cogniƒve work)

Commercial Peer-toPeer or Peer-centred two sided plaorms mostly unregulated

NFP=noorprofit;FP=forprofit; BL=businesses-led;PC=peer-centred

High skilled labour intensive online service (non- rouƒne cogniƒve work)

Renƒng space or car, selling goods

rides haring Low skilled labour intensive localised service

A= assets (property, money, skills); T=free ƒme (bare labour); P2B=peers to businesses; P2P=peers to peers

Conƒnuum

Empirical marginal type

BL

A

T

PC

Fig. 4:  Heuristic Conceptual Mapping of a Sharing Economy. Source: Re-elaboration from Codagnone et al. (2016b, pp. 22–25). This all-encompassing definition contains the elements upon which a heuristic typology with some relevant distinctions could be built. It is a two-step conceptualisation graphically illustrated by the picture below. Obviously, as any conceptualisation, it represents a simplification over the nuances that one can detect empirically, but it works fairly well for our purpose of identifying those platforms that are the centre of disputes and policy concerns and for making further distinctions even within this group. In the left-hand side of Fig. 4, platforms are categorised and distinguished according to their commercial orientation (for profit vs. NFP) and interaction modality (in the sense of whether the transaction is peer-centred/led or businesscentred/led). Another element that cannot be rendered in a 2×2 typology is the dimensional relevance11 of platform that we comment discursively. These three dimensions shape regulatory and policy relevance of platforms. Hence, the first typology in the bottom left of Fig. 4 basically identifies one group as the most relevant one for both analysis of rhetorics and regulatory and policy issues. This is the Southeast quadrant where we have the largest commercial platforms (i.e. Uber, Airbnb, TaskRabbit and Upwork) that are allegedly (see discussion of control in Chapter 1 and again in Chapter 3) P2Ps or P2Bs and to date mostly unregulated and at times object of legal disputes. They have a 11

Some have a users’ base of a few hundreds or thousands of individuals and others of millions of people. In terms of dimensional relevance, for instance, makers’ spaces and collaborative production platforms are currently much smaller than transactional (i.e. Airbnb) and labour (i.e. Upwork or Freelancers) platforms; the latter are a more immediate source of regulatory concerns, whereas innovation policy support measures may enable the former to scale up.

44    Platform Economics large user base, raise short-term regulatory concerns (market access, taxation, consumer protection and liability and labour laws) and some of them disrupt incumbent industries triggering their reaction. The other three types are not relevant for our analysis, hence not considered. This most relevant group still includes platforms with important differences,12 and thus is further broken down into four types in the typology placed on the top right part of the figure. In this case, we again use the interaction modality that can broadly take two discrete values: exchanges between peers (P2P) and peers selling their labour to businesses (P2B). In this case there is a difference between an individual delivering service to another individual (i.e. Task Airbnb) and an individual delivering service to a firm or to another individual who use that service for an entrepreneurial activity (i.e. Upwork or Amazon Mechanical Turk). In the former case, the exchange concerns more consumer issues whereas in the latter, the exchange concerns labour issues. There are, however, cases that have both implications as labour-intensive services delivered to final consumers (i.e. Uber). The second dimension must be seen instead as a continuum and pertains to those that have been framed as idle assets that sharing economy brings back to live. This dimension can range from tangible assets such as capital (money, property, real estate, although in some cases people put on platforms space that they rent and do not own) or durable goods (cars) to less tangible assets, such as skills or just the free time, to offer very low or no skill labour-intensive services. Obviously, letting an apartment on Airbnb and a car with Relay Rides, or selling goods, also requires some work. It is equally obvious, however, that in these cases most or the larger part of the value obtained comes from property or goods unlike doing errands for other peers in TaskRabbit or performing micro-tasks for businesses in Amazon Mechanical Turk. In the northwest quadrant (Quadrant 1), the extreme ideal/typical case is the lending of money with basically no labour input from individuals to small businesses (i.e. Funding Circle). Since lending (and to some extent crowdfunding platforms) represents vertical specific activities with very distinct regulatory implications, they are no longer considered here. On the other hand, considering skills as an asset as compared to just free time for low skill jobs, this quadrant identifies platforms that digitally match skilled labour-intensive services to fully delivered businesses online. Next, in the northeast quadrant (Quadrant 2), we have more or less asset-intensive provision of services and selling of goods. The southeast quadrant (Quadrant 3) includes platform-matching peers providing labour-intensive services to other peers (final consumers) through unskilled manual work (i.e. TaskRabbit). Uber somehow 12

For instance, ride services (i.e. Uber), as compared to ride sharing (i.e. BlaBlaCar) and car sharing (i.e. RelayRides), differ in several ways: (1) Ride services are labour-intensive, whereas car- and ride-sharing are not; (2) ride services raise issue of market access (i.e. licensing) and provoked strong protests from incumbents, whereas car- and ride-sharing don’t; and (3) car- and ride-sharing create much less problems for what concerns liability and insurance. All of these transportation-centred services have obviously different implications compared to space rental or re-sale of goods. Digital labour platforms in turn present their own peculiarities.

Labour Intermediation Platforms    45 separated because, while it entails unskilled manual work from the drivers, this activity is enabled by a tangible asset. Finally, the southwest quadrant (Quadrant 4) covers ‘labour-intensive unskilled provision of services to businesses’. All platforms, mostly based on labour, are further discussed in Chapter 3. In Quadrant 2, we find platforms raising regulatory concerns regarding consumer protection. In this quadrant, Airbnb has also been the object of other controversies (zoning, taxes and local rules for short-term rental). On the other hand, Quadrants 1, 3 and 4 have implications for employment and social protection that are not relevant for Quadrant 1. On the other hand, ride-sharing (Uber) is the utmost source of concern, as it entails both consumer- and labour-protection issues. In this respect, this typology helps to appreciate different regulatory implications of ride services (Uber) as compared to ride-sharing (i.e. BlaBlaCar) and car-sharing (i.e. RelayRides). The former is labour-intensive and currently at the centre of labour disputes, whereas car-sharing entails little or no work, and ride-sharing only a limited amount of work. In P2P car- and ride-sharing, reservations are made in advance, the two peers eventually meet, and driving is mostly for personal use, with less frequent but longer utilisation. In ride services, on the other hand, scheduling is on demand with a short lead-time, the driving is for commercial use and utilisation is very frequent (with more risks entailed). When one peer is the owner, just giving a ride or renting his/ her car, and not a driver carrying paying passengers, liability policy is much more straightforward. The fact is that provider and consumer meetings increase trust; and the less frequent use reduces risks to safety. From a regulatory perspective, these are important factors. Last but not least, as shown later, Uber is possibly the only ‘sharing’ platform which could become the object of concern regarding competition laws.

How Big is the Sharing Economy? We could simply close this section answering: ‘We don’t know’. Estimates are almost as floating and inconsistent as terminological usage. According to the UK Office for National Statistics (ONS, 2016), lack of a common definition and understanding of the ‘sharing economy’ is the main obstacle to measure ‘sharing economy’ in terms of either its economic value or the number of individuals involved as users or providers. With regard to the first measurement (monetary), the ‘sharing economy’ does not fall within standard classifications used in business and economic statistics. The volume of P2P transactions is almost entirely lost to consumer statistics and price indexes. As regards measurement through surveys, the ONS reports the result of piloting qualitative interviews and casts doubts on the reliability of these surveys. Individuals and experts interviewed, or those who participate in focus groups, have very different understandings of what the ‘sharing economy’ is. Those who have used ‘sharing’ platforms have clear difficulties in recalling exactly how many times they did it, the average expenses they incurred or the income they obtained. Thus, these aspects cast some doubts on the data coming from surveys of general populations. Bearing these caveats in mind, some exemplificative estimates of monetary values and findings from surveys are reported below.

46    Platform Economics The first global estimate dated 2013 put the revenues accruing to participants on sharing platforms at $3.5 billion (Geron, 2013). A report released by PwC in August 2014 calculated that on a global basis the ‘sharing economy’ was worth $15 billion and could reach $335 billion by 2025 (Vaughan & Hawksworth, 2014).13 A study released in January 2016 by the French government estimated that in France the ‘collaborative economy’ activities turnover was $2.5 billion, involving about 15,000 firms (including self-employed micro-entrepreneurs) and generated 13,000 permanent jobs (Barbezieux & Herody, 2016). This would amount to approximately 0.1% of French GDP generated by 0.5% French companies for 0.05% of French total employment. A study released in 2016 by the European Parliament estimated that in 2015 the value the sharing economy was €20 billion and could grow up to €572 billion by 2025. Later, in 2017, the European Parliament Press Room released these new estimates14: The value of all transactions mediated by ‘collaborative platforms’ in EU28 went from €15.9 billion in 2014 to €28.1 billion in 2015 whereas revenues to platforms increased from €1.8 billion in 2014 to €3.6 billion in 2015; online labour markets were reported as intermediating in 2015 transaction for a value of €2.7 billion. A study commissioned by DG Justice of the European Commission has estimated the following figures using data from surveys on making a number of assumptions applied to underlying statistics: (a) Total annual consumer spending in online P2P platforms in EU28 (around 2017): €27.9 billion (€17.8 billion resale of goods; €6.6 billion accommodation renting; rides: €1 billion; odd jobs: €1.2 billion; other renting: €1.3 billion); (b) total revenues for peer providers in EU28 (around 2017): €17.2 billion (€10.8 billion resale of goods; €4.1 billion accommodation renting; rides: €794 million; odd jobs: €763 million; other renting: €820 billion; (c) revenue to platforms in EU28 (around 2017): an uncertain cut (about €4 billion) of about €10 billion that represent the difference between consumer spending (€27.9 billion) and €17.2 billion of revenues accruing to providers (Hausemer et al., 2017, pp. 41–52). Another study commissioned by the Commission to PwC estimated that: (a) household tasks mediated through platforms in Belgium, France, Germany, Italy, the Netherlands, Poland, Spain, Sweden and the UK totalled a value of about €2 billion with €450 million accruing to platforms; and (b) online professional task platform-mediated transactions worth €750 million with platforms’ revenue of €100 million (Vaughan & Daverio, 2016). In 2016, considering 173 online labour platforms and their active providers in EU28, it was estimated that platforms’ revenue amounted to about €4.2 billion, which is equivalent to 13

This estimate by PwC has been spun around in the last two years and reified as a true ‘quantification’. Few of those citing this estimate bothered, as the authors of this essay did, to dig into how these figures were constructed to discover they are based on a shaky methodology and a controversial inclusion of sectors and players, such as, for instance, Netflix and Spotify. Netflix and Spotify are resellers and have nothing in common with platforms, such as Airbnb or Uber, classically typifying the ‘collaborative’ or ‘sharing economy’. 14 http://www.europarl.europa.eu/news/en/headlines/economy/20170428STO72971/ infographic-the-increasing-popularity-of-the-collaborative-economy

Labour Intermediation Platforms    47 0.03% of EU GDP (De Groen, Kilhoffer, Lenaerts, & Salez, 2017, p. 350). As visible, this above sample of estimates providing variable numbers, using different definitions, is reported just for the sake of illustration. Most of these estimates should be taken with caution because of the lack of reliable data and consolidated empirical evidence and are inevitably based on questionable assumptions. It is certainly a statistically detectable and non-marginal phenomenon, but the extent of this quantitative significance is still far from being quantified in robust and reliable ways. The first large survey conducted in 2013 (including the UK, the US and Canada), reports that 29% of the British population had engaged at least once in a ‘sharing’ transaction and 23% used one or more platforms such as Airbnb, Uber, TaskRabbit, Etsy and Kickstarter (Owyang, Samuel, & Grenville, 2014). A report by the National Endowment for Science, Technology and the Arts (NESTA) estimated that in 2014, 25% of the UK adult population shared online in some way (Stokes, Clarence, Anderson, & Rinne, 2014). A survey based on a nationally representative sample of the UK population aged 16–75 years and conducted at the end of 2015 reported that 72% of the respondents are either making an income from online activities or buying labour from others (Huws & Joyce, 2016b). Of these, around 1% is involved in online rental schemes such as Airbnb. Another survey by the same authors conducted in Sweden at the beginning of 2016 provides similar shares (Huws & Joyce, 2016a). Around 68% of the Swedish adult population are active in some way in online economy, for instance, selling goods online or renting out rooms on platforms such as Airbnb (those involved only in online rental are around 1% in Sweden as well). A large representative survey of Amsterdam’s citizens shows that 38% of respondents are willing to take part in all possible forms of ‘collaborative consumption’ and 84.1% are willing to take part in at least one form (van de Glind, 2013). Another study by the French government reports the findings of a survey conducted in 2009, indicating that 89% of the respondents had engaged at least once in a ‘collaborative consumption practice’ (PIPAME, 2015). In Denmark, the national statistical office has included a short module on the sharing economy. It focuses on ride services and renting in the traditional ICT usage by household survey. It found that by mid2015: (a) 3.1% of Internet users let out through digital platforms and 8.7% rented from Airbnb and similar platforms abroad and 4.4% in Denmark; (b) Uber was used by 2.8% of Internet users (Nielsen, 2015). One ad hoc edition of the Eurobarometer reports that in 2016 more than half of the respondents have heard of collaborative platforms (52%) and around 17% of respondents say that they have used them (European Commission, 2016d). Other highlights from the Eurobarometer include the following: younger and more highly educated respondents who live in more urban areas and who are self-employed or employees are much more likely than the average citizen to be aware of collaborative platforms (63%) and to have used the services of these platforms at least once (32%). Over onethird of the respondents who have visited collaborative platforms say that they have provided services on these platforms (32% of 17% respondents). Almost one in 10 respondents who have visited collaborative platforms have provided services on these platforms once (9%), while almost one in five of these respondents offer

48    Platform Economics services via this type of platforms occasionally – once every few months (18%). Finally, one in 20 says that they offer services via these platforms regularly – every month (5%). As for the case of estimates of monetary values, surveys also suggest that the phenomenon is statistically detectable and non-marginal, but the precise level of participation remains an empirical question.

Rhetorical Discourses There are several contributions that explicitly or to a large extent focus on rhetorical and discourse analysis (Banning, 2016; Belk, 2014a, 2014b, 2016; Cohen & Muñoz, 2015; Dredge & Gyimóthy, 2015; Fitzmaurice et al., 2018; Gruszka, 2017; John, 2013a, 2013b; Kennedy, 2016; Lee, 2015; Martin, 2016; Ravenelle, 2017; Richardson, 2015; Schor, 2014, 2015; Schor & Attwood-Charles, 2017; Walker, 2015). Digital sharing has been characterised, for instance, as info-liberalism, a paradigm in which online sharing becomes entangled and aligned with neo-liberal capitalism despite its contradictory semantics (Banning, 2016). Banning shows how the positively value-loaded discourse of sharing is obliquely exploited to fuel the ‘algorithmically regulated economy’ of sharing platforms, which affect prime users through algorithms. According to Kennedy (2016), platforms have cultivated rhetoric of sharing in as expression of a networked culture appropriating normative discourses on community, generosity, shared values of cooperation and participation. Based on in-depth interviews with active participants in sharing platform in the San Francisco area, Cockayne (2016) argues that the sharing discourse has maintained an emotional association to issues such as community, inclusion and participation. This occurs despite the fact that in practice many ‘sharing platforms’ are commercial, function based on digital peer review and punitive rating systems, whilst they are algorithmically mediated, precarious, and indulge in ‘entrepreneurial’ contract work. The author illustrates how sharing has been harnessed to justify and normalise flexible and precarious work through a paradoxical alignment between capitalist exchange and altruistic social values while recognising that the sharing economy is contrasted with harsh criticism considering platforms as quintessential players of neoliberal capitalism nonetheless. A qualitative study of sharing practitioner in Vienna identified four possible framings: visionary supporters, market optimists, visionary critics and sceptics, ‘each bringing their values, visions and practical goals characteristic of different understanding of the collaborative economy’ (Gruszka, 2017). The author concludes that, in the face of polarisation between supporters and critics, it is questionable to use a globally applicable umbrella definition of sharing or collaborative economy and that more local and contextual explorations of practices at local level would prove fruitful. The sharing or ‘gig’ economy claims to bring the romance of entrepreneurialism to the masses. Through P2P technology, workers can monetise their homes, resources, time and skills to make additional money. What is marketed as an empowering business opportunity is laden with difficulties and contradictions. Sudden changes to platform design, service offerings and algorithms leave workers feeling vulnerable, not independent. Instead of

Labour Intermediation Platforms    49 embracing sharing economy rhetoric, most workers describe themselves as simply seeking money. This article sheds light on the diversity of the gig economy and questions sharing economy companies’ claims that they are contributing to the growth of entrepreneurship (Ravenelle, 2017). In both Dredge and Gyimóthy (2015) and Martin (2016) one can find attempts at systematising rhetorical discourse. Dredge and Gyimóthy (2015) summarise the rhetorical field as comprising five themes about social technologies: unlocking hidden wealth; enabling more equal distribution of benefits; facilitating resilient communities, authentic relations and the moral economy; invoking the invisible hand; and making self-regulation possible and effective. These authors further argue that according to the discourse analysis presented, framing the sharing economy has gone through four stages: the book and Ted stated by Botsman (2010) brought the topic to wider attention; next came the powerful message that technologically enabled matching and reputational trust could unlock idle assets, favour the activation of the commons and trust among strangers and lead to sustainable consumption; this was followed by large-scale public relation campaigns by Uber, Airbnb and other platforms which established research departments and government relations department and flooded the debate with their own reports (based on opaque and unreliable methodologies) about their benefits; fourth stage, proliferation of discourses has been delegated to advocacy groups (Dredge & Gyimóthy, 2015, pp. 3–5). On the other hand, Martin (2016) identifies six discourses (three positives and three negatives): economic opportunity; more sustainable form of consumption; path to decentralised and equitable economy; unregulated market places; reinforcing neo-liberal paradigm; and incoherent field of innovation. Despite the fact that several authors have challenged and deconstructed the rhetoric about community and the moral economy, Fitzmaurice et al. (2018), considering the perspectives of 120 participants interviewed in depth, insist on the moral economy dimension and the potential for domesticating and taming this market. They report that these participants (providers in two for-profit and three NFP sites) see the sharing economy as ‘an opportunity to build a radically different market, from the bottom up’. Other sources, among those reviewed, are instead themselves a source of normative and rhetorical narratives of both an overly optimistic and an overly pessimistic nature (Agyeman McLaren, & Schaefer-Borrego, 2013; Allen & Berg, 2014; Caldararo, 2014; Cohen & Sundararajan, 2015; Guttentag, 2013; Heimans & Timms, 2014; Heinrichs, 2013; Koopman et al., 2014, 2015; Kuttner, 2013; Matzler & Kathan, 2015; Morgan & Kuch, 2015; O’Regan, 2009; Sundararajan, 2014; Thierer et al., 2015; Wittel, 2011; World Economic Forum (WEF), 2013, 2014; Wosskow, 2014). From these sources and some of their antecedents, one can identify the following three broad narratives: (1) social utopianism,15 15

The roots of social utopianism can be found in the popularisation by Howe (2006, 2008) of the ‘wisdom of crowds’ narrative (Surowiecki, 2004) and in the well-known narrative on the creativity of the commons (Benkler, 2004, 2006). In this optimistic vein, one can also find the democratising effect of the ‘long tail’ (Anderson, 2006), the generosity stemming from ‘cognitive surplus’ (Shirky, 2010), the celebration of

50    Platform Economics (2) business optimism16 and economics laissez faire17 and (3) social pessimism.18 Each can be associated with corresponding grand narratives about the future: great transformation (social utopianism), growth-oriented globalisation (business optimism and economic laissez faire) and barbarisation or uberisation (social pessimism). The first is a vision of change entirely led by community, whereby the social re-embedding of the economy is achieved entirely because of changes in behaviour and culture. In the second, individuals and innovative firms are empowered in a more competitive and individualistic way, market forces are left uncontrolled. The third sees firms and work as being dis-intermediated, decentralised and deconstructed into smaller elements, to be re-intermediated through algorithmically enabled panoptical control. It is the world of ‘lumpen cognitariat’ or ‘algorithmic salariat’ and ‘algocracy’. Workers are substituted by robots or are transformed into ‘robots’ (i.e. mechanical Turks), performing highly routinised and repetitive micro tasks. Dis-embedding and dis-empowerment increase unemployment and inequality to unprecedented levels and further fuel antagonistic feelings not only towards ‘out-groups’ (i.e., immigrants) but also towards ‘in-groups’. From the above analysis and discussion, we singled out five themes that we further develop in the following sections: (1) neo-liberal co-optation of the ‘sharing’ rhetoric and movement; (2) social capital and community revival; (3) distributional and stratification effects (i.e. inequality, race discrimination and labour issues); (4) environmental (greener commerce, less carbon foot-print and CO2 crowdsourcing as a model for problem-solving (Brabham, 2008, 2013; Gehl, 2011), the philosophical praise of its virtues (Benkler & Nissenbaum, 2006) and the promise of increased efficiency (Chandler & Kapelner, 2013; Djelassi & Decoopman, 2013; Satzger, Psaier, Schall, & Dustdar, 2013). Other narratives more directly related to the sharing movement talks about the triple win of greener commerce, greater profits and rich social experiences such as community revival and strengthening of social capital (Leadbeater, 2009; Grassmuck, 2012b; O’Regan, 2009; Wittel, 2011). Anthropological and neuroscience contributions are cited, for instance, to argue that sharing is an evolutionary and cultural traits of human beings (Agyeman et al., 2013). 16 A number of sources present a sort of business-driven optimism (Guttentag, 2013; Heimans & Timms, 2014; Matzler & Kathan, 2015; WEF, 2013, 2014; Wosskow, 2014). Management gurus (Heimans & Timms, 2014), for instance, propose a normative loaded and speculative distinction between ‘new power’ (sharing economy, but also grassroots political movements) and ‘old power’ (big corporations, but also established political parties). 17 Neo-liberal and libertarian economists present the usual narrative on free market and self-regulation (Allen & Berg, 2014; Cohen & Sundararajan, 2015; Koopman et al., 2014, 2015; Sundararajan, 2014; Thierer et al., 2015). They expect the ‘sharing’ platforms to: (a) increase economic activities and productivity and create new jobs by leveraging ‘dead capital’ and lowering; and (b) reduce information asymmetry between consumers and producers, thanks to reputational ratings that as a form of selfregulation make traditional regulation redundant. 18 Very critical and pessimistic observers have defined crowd-employment platforms as the new sweatshops (Uddin, 2012; Zittrain, 2009), and analysed them as new forms of encroachment and exploitation of labour (Carr, 2008; Deuze, 2007), underpaid free work (Kleemann, Voß, & Rieder, 2008; Scholz, 2013) and of new digitally enabled surveillance (Aneesh, 2009).

Labour Intermediation Platforms    51 emissions) and socio-economic impacts (consumers welfare, efficiency gains, impacts on disrupted industries and (5) regulation (i.e. ratings as forms of selfregulation). The choice of these themes does not have the ambition of presenting an exhaustive systematisation of all possible rhetorical discourses. Its rationale is two-fold: We selected themes that could be contrasted with available empirical evidence and/or are much debated and relevant from a regulatory and policy perspective.

Lobbying as Framing: Harnessing Rhetoric and Evidence The first critical issue concerns the neo-liberal co-optation of the ‘sharing’ rhetoric and movement, which is a case in point of how rhetoric and evidences are used to frame policy debate and exert lobbying. This controversy divides the activists of the ‘sharing movement’ into ‘what is’ and ‘what is not’ the ‘sharing economy’ (Belk, 2014a, 2014b; John, 2013a, 2013b). When an ethnographic study (Bardhi & Eckhardt, 2012) produced evidence that Zipcar members were not inspired by a sense of community and altruism, this was vehemently rebuffed in the online magazine Shareable. When in 2010, CouchSurfing became a for-profit ‘B Corporation’,19 it gave rise to controversies and heated debates. These examples testify to the value-loaded ways in which sharing was originally conceived. Activists and hippies supporting CouchSurfing are supposedly in a different league from the founders and users of Airbnb. It further demonstrates the rhetorical and ambiguous significance and importance that the vocabulary of sharing has assumed in the context of ‘sharing economy’. Since sharing has a positive and progressive connotation, more and more companies have started to claim that they are part of ‘sharing economy’. In this respect, it argumented that large companies have co-opted the sharing movement to pursue their own economic interests through traditional lobbying strategies (Lee, 2015; Schor, 2015; Walker, 2015).20 In this respect, it has been suggested that Silicon Valley is the new revolving door for Obama staffers with much emphasis placed on the fact that Uber appointed former Obama campaign manager David Plouffe as chief of policy and strategy (Kang & Eilperin, 2015) and provided its data to Alan Krueger, former chairman of President Barack Obama’s Council of Economic Advisers, to produce a controversial paper concerning impacts on drivers’ earnings (Hall & Krueger, 2015, see infra). According to Lee (2015, p. 17), the ‘sharing economy’ is just another example of how ‘insurgent sentiments’ are used to ‘sell the bona fide

19

A ‘B Corp is to business what Fair Trade certification is to coffee or USDA Organic certification is to milk (https://www.bcorporation.net/what-are-b-corps, accessed on June 8, 2015). 20 In this respect, it has been suggested that Silicon Valley is the new revolving door for Obama staffers with much emphasis placed on the fact that Uber appointed former Obama campaign manager David Plouffe as chief of policy and strategy (Kang & Eilperin, 2015) and provided its data to Alan Krueger – the former chairman of President Barack Obama’s Council of Economic Advisers – to produce a paper concerning impacts on labour matters.

52    Platform Economics of profit-making corporations’. The anti-establishment ideology disseminated by Sharable and the association Peer.org are increasingly seen as mouthpieces of big companies such as Uber and Airbnb that use this rhetorical weaponry to pursue their own economic interests (Kerr, 2014). In a Harvard Business Review piece, Cannon & Summers (2014) advised big players in the ‘sharing economy’ to lobby in various ways, using for example reports about their positive impacts, in order to frame potential regulation as hampering, not only economic innovation, but also the trickled down of benefits to society, and especially to the middle and lower class consumers. Indeed, both Airbnb and Uber (and other larger players) are pursuing aggressive market growth strategies and calling for no regulation in a classical neo-liberal fashion. Moreover, it seems that they have followed or anticipated this advice, cobbling together evidence to show the benefits that are produced by their products/services. Few young tech start-ups have publicly ruminated on their economic impact in quite the same way as done by Airbnb (Badger, 2014). Uber has also released reports on the benefits of its services. These are summarised in Boxes 1 and 2. At the same time, they are exploiting the ‘sharing’ rhetoric in public relations activities: they are participating in official hearings and releasing their own reports on their positive social impacts that they have. In a public hearing with the UK House of Lords (2016), Patrick Robinson, Head of Public Policy Europe and Canada for Airbnb, affirmed as follows: In our case, the public interest at stake here is, first, about consumers and consumer choice not just to consume services but to be producers of services too. The additional income that Airbnb hosts are making is very important to them. Identifying outdated rules and regulations that might stop people engaging in what is beneficial activity is a good exercise and one that I am delighted that we undertook in London earlier this year when they introduced the new rules allowing people to share their homes without the need to seek planning permission, for example. That then raises other issues that we need to be mindful of. We need always to be mindful of the impact that that could have on neighbours or on local communities. That is why we and many other companies in the collaborative economy spend a lot of time measuring impacts and finding effective ways to deliver good public interest outcomes. (p. 45) Besides these reports published in their blogs, Airbnb and especially Uber have recruited prominent academics and/or former member of the presidential administration to write papers about their positive impacts. Airbnb commissioned ­former White House National Economic Advisor, Gene Sperling (2015) to write a report, which allegedly shows that the platform creates a 14% annual increase ($7,350 per year for an average of 66 days of hosting) in the income of middleclass families hosting guests at their homes. The first Uber paper of this kind was written by Alan Krueger, prominent academic and former member of the Obama administration, and J. Hall, a researcher

Labour Intermediation Platforms    53

Box 1: Airbnb Self-reported Impacts. ⦁⦁ 81% of hosts share the home where they live, 52% earn low to moderate

⦁⦁ ⦁⦁

⦁⦁ ⦁⦁

incomes, 53% affirmed that hosting helped them remain in their home, and 48% that they use earnings from hosting to pay for regular household expenses. Conducted in 2012, the study on S. Francisco found that Airbnb generated approximately $56 million in local spending and supported 430 jobs; A 2013 study reports that in Paris, Airbnb generated €185 million (approximately US$240 million) of economic activity in Paris, and supported 1,100 jobs; Conducted in 2014, the study on the UK reports that Airbnb generated £824 million in economic activity and supported 11,600 jobs; Other studies report that Airbnb supports 1,600 jobs in Sydney and 4,000 in Barcelona. Source: Airbnb (2015b).21

Box 2: Uber’s Self-reported Impacts. ⦁⦁ Uber has published various reports on its impacts on cities. These are pre-

sented as being generated by additional rides (in addition to those of the established taxi industry) and creation of jobs (Uber, 2014, 2015d, 2015e). ⦁⦁ For Chicago, for instance, $43 million from new economic activities, and 1,000 new jobs are reported. In another report, the ‘sharing economy’ giant claims that its services are improving the reliability of drivers in communities where military bases are located (Uber, 2015b). ⦁⦁ It also claims that Uber has contributed to a sharp decline in the Driving Under Influence of alcohol or drugs (DUI) rate in cities where it operates. ⦁⦁ The most ambitious effort is a two-volume report entitled Ubernomics: How Ridesharing Can Impact The German Economy (Uber, 2015f, 2015g). This report argues that with Uber, more rides and lower prices will increase consumer welfare and provide more earnings to drivers. In order to produce this study, Uber collaborated with Justus Haucap, director of the Düsseldorfer Institut für Wettbewerbsökonomie (DICE) and former chairman of the German Monopolies Commission, and with DIW Econ, the consulting arm of the German Institute of Economic Research (DIW). Sources: Uber (2014, 2015a, 2015b, 2015c, 2015d, 2015e, 2015f, 2015g).

21

From this entry in the blog summarising the overall results, one can then proceed to access the city-specific reports (Airbnb, 2012, 2013a, 2013b, 2013c, 2013d, 2014a, 2014b, 2015a, 2015b; Uber, 2014, 2015a, 2015b, 2015c, 2015d, 2015e, 2015f, 2015g).

54    Platform Economics at Uber (Hall & Krueger, 2015). The paper, which did not go through peer-review screening of the dataset used, claimed that Uber drivers’ per hour earnings were higher than those of regular cab drivers. This soon created a controversy, as investigative journalistic pieces showed that the claim was false, for it did not factor in expenses. In January 2017, Uber settled with the Federal Trade Commission on the grounds that they had made deceptive claims about drivers’ incomes (Schor & Attwood-Charles, 2017, pp. 7–8). Many other of these papers about Uber’s positive impacts have been written by the privileged few, to which the company gave access to sections of its data (Angrist, Caldwell, & Hall, 2017; Castillo, Knoepfle, & Weyl, 2018; Chen, Chevalier, Rossi, & Oehlsen, 2017; Cohen et al., 2016; Cook et al., 2018; Hall, Horton, & Knoepfle, 2017; Landier, Szomoru, & Thesmar, 2016; Rizk, 2017), although which company data was shared is only known to these authors, and yet remains unknown to the rest of the academic communities and to the general public. These papers treat different aspects in different localities, but they have the following two characteristics in common: (a) they have not gone through peer review and (b) among the authors, along with well-known economists there were always one or more Uber employees. The only paper based on Uber data that has been published in a peer-review journal is a study on the efficiency of ride-sharing services vis-à-vis taxis by comparing the capacity utilisation rate of UberX drivers with that of traditional taxi drivers in five cities (Cramer & Krueger, 2016). We elaborate further on this in Chapter 3.

Social Capital and Motivation The evidence on the motivation of individuals to participate in sharing activities and on the impact that this may have on social capital22 and generalised trust is

22

The concept of social capital commands an ever-expanding body of literature that cannot be discussed here. General approaches define social capital in slightly different ways depending on the theoretical perspective (Bourdieu, 1986; Coleman, 1988, 1990; Putnam, 1993, 2000). At a very basic level it can be said that the concept entails both normative (norms and values) and instrumental dimensions (networks). At macro level, social capital can be equated to civic sense entailing norms, social values, trust and social network (especially participation in association). At a more microlevel, social capital can be defined as the ensemble of social networks that can enable individuals to gain access to desired resources and outcomes. It is, however, most important for our purposes here to consider a bit further the concept of trust and relate it to social capital, as trust is fundamental for the uptake of any sharing economy platform. Trust is the social glue that enables collaborative consumption marketplaces and the sharing economy to function without friction. Scholars of trust distinguish between generalised and particularised trust (Couch & Jones, 1997; Delhey et al., 2011; Freitag & Traunmüller, 2009; Putnam, 1993, 2000; Stolle, 2002; Yamagishi & Yamagishi, 1994; Yamagishi et al., 1998). Particularised trust, also referred to as ‘thick trust’ (Putnam 2000), concerns a close network of social proximity (i.e. family and friends). Obviously, since for sharing economy platforms to scale up transactions among strangers are crucial, generalised trust as a willingness to rely on ‘abstract others’ is crucial. As shown later, however, also particularised trust can matter as a result of platforms enabling users to see what their ‘friends’ do and how they rate their experiences.

Labour Intermediation Platforms    55 at best mixed, and certainly not fully supportive of the narrative on the revival of community and social ties that would be brought about by sharing platforms. According to a review by Schor and Attwood-Charles (2017, p. 6), the dominant motivations to use sharing platforms are financial and not social. An online survey, conducted on a convenience sample of 800 tourists having used the platform in 2015, finds that respondents were most strongly attracted to Airbnb for its practical attributes, and somewhat less so for its experiential attributes (Guttentag, Smith, Potwarka, & Havitz, 2017). On the other hand, based on qualitative and quantitative data and using a scale of sharing motives, another study reports that reasons for participating in online sharing platforms are more nuanced than previously thought (Bucher, Fieseler, & Lutz, 2016). The authors conclude that sharing motives are driven by moral, social-hedonic and monetary aspects. A survey testing the extent to which participation in one ‘sharing’ platform was a form of ‘anti-capitalism’, found that individuals have different motives and identified four clusters: ‘socialites’, ‘market avoiders’ and two other profiles with no particular ideological motivation (Ozanne & Ballantine, 2010). Lamberton and Rose (2012) found the same mix of utilitarian and socially/environmentally oriented motivations in a study based on three surveys of the users of three different platforms. A qualitative study of time banks found that anti-capitalist sentiments, discontent with consumption and an ideology of sustainability emerged as strong motivations for participation (Dubois et al., 2014). The authors, however, also found that the different levels of cultural capital and the distinctions they produce matter and create contradictions.23 An ethnographic study of Italian home-swappers (Forno & Garibaldi, 2015) also found both social and utilitarian motives. Sustainability, enjoyment of the activity and economic gain were the key motivations found among users of a platform in Finland (Hamari et al., 2016). A large-scale survey of free reuse groups (e.g. Freecycle and Freegle) in the UK shows that the majority of participants do have significantly stronger self-transcendence (i.e. pro-social) values than the wider UK population, but they also have other more extrinsic values (Martin & Upham, 2016). Möhlmann (2015) found a mixture of self-interest and socially oriented motivation through surveys of users of the car-sharing service Car2go and Airbnb (N = 187) in Germany. The Baumeister and Wangenheim (2014) survey of a representative sample of 2,000 German respondents, randomly assigned to express their views and attitudes to accessing rather than owning different types of products, found that the attitude to access is consistently worse than the attitude to ownership across all product categories. An in-depth qualitative study of Freecycle found thick relations and social capital at work and also

On the other hand, the optimistic expectations about the sharing economy were that it would increase social capital in both the forms of trusting others and of participating in ‘community’-based activities, and that it would create new meaningful friendships and social experiences. 23 Cultural capital and distinctions are used by the authors in the sense as specified by Bourdieu (1984, 1986).

56    Platform Economics tensions between the goals of the institution (the owners of the Freecycle brand) and those of its community members (Arsel & Dobsha, 2011). Three exploratory studies of local-level platforms found that while traditionally relational and reciprocal exchange is highly valued, weak ties of non-reciprocal exchange allow the communities to tap into the significant distributed expertise of community (Ozanne & Ozanne, 2011). A qualitative empirical analysis of non-monetary market places (Really Really Free Markets), which blend online and offline sharing events, found that a sense of community is both a driver of participation and an outcome of these events (Albinsson & Yasanthi Perera, 2012). Findings of the ethnographic study done by Bardhi and Eckhardt (2012) on Zipcar users came as a thunderbolt for both activists and earlier scholars of the ‘sharing economy’. The authors report that Zipcar members did not feel any sense of attachment to the organisation, as their main motivation was use value. These consumers did not refer to hedonist or altruistic values, and engaged in opportunistic behaviour towards the company and one another (negative reciprocity). An empirical qualitative analysis of gift-giving, sharing and commodity exchange at Bookcrossing.com underscored the importance of collective reciprocity and anonymous sharing (Corciolani & Dalli, 2014). Two studies done by Parigi and colleagues on social capital and social networks focussing on CouchSurfing also yielded mixed results (Parigi & State, 2014; Parigi et al., 2013). The first study consisted of a network analysis using data obtained by CouchSurfing.com for the period 2003–2010 (Parigi et al., 2013). A random sample of 10,000 American users’ monthly logins were counted and recorded over their career, in order to analyse how they associated with each other. The study tested the following two alternative hypotheses about individuals’ participation in associations: (1) participation as a by-product of existing friendships; (2) participation driven by the association’s capacity to form new identities. The authors reported that new friendship ties had a significant impact on participation, whereas pre-existing ties (defined here as ties with other members formed outside organisation’s context) had a negligible impact. This would seem to suggest that a platform such as CouchSurfing is generating new social capital. On the other hand, the second study, where quantitative analysis was integrated with ethnographic work, produced a somewhat paradoxical result on the disenchanting effect of technology (Parigi & State, 2014). The accumulation of ratings about users (whether guests or hosts) had a double-edged effect on the emergence of trust and relationships; it made relationships easier to establish initially, but it also weakened them above a certain threshold. That is to say, technology facilitated the emergence of interpersonal trust among CouchSurfers but it also made establishing of strong ties harder as users acquired more and more reviews. This case illustrates a process of disenchantment created by technology, where technology increases the ease with which friendships are formed and, at the same time, diminishes the bonding power of these experiences. It emerges from the reviewed evidence that (a) the motivations that lead individuals to join the ‘sharing economy’ range from altruism to utilitarian goals and

Labour Intermediation Platforms    57 also include a scattering of anti-capitalist and anti-consumption ideologies and sentiments; (b) the ‘sharing economy’ creates some form of genuine social capital but is also based on reciprocal (negative and positive) exchanges; (c) judging from the reviewed sources, altruistic and ideological motivations and social capitalbuilding clearly seem to characterise more the early NFP initiatives. It can be stated that, going beyond the polarised rhetoric and controversies, the ‘sharing economy’ overall is a mixture of ‘passions’ and ‘interests’.

Distributional and Stratification Effects The third rhetorical theme concerns the claim by enthusiasts and platforms themselves that the sharing economy provides more access and opportunities, and reduces inequality and discrimination. Whether platforms improve, transform or worsen inequalities based on class, race or position in the labour market (this topic is anticipated here and discussed more in depth in Chapter 3) is hotly debated and there is little empirical evidence supporting the claims of platforms. One place to start is the earlier cited Sperling (2015) report commissioned by Airbnb; it states that that the platform creates a 14% annual increase ($7,350 per year for an average of 66 days of hosting) in the income of middle-class families. In this respect, two observations are in order: As there are currently about two million listings worldwide and some individuals have multiple listings, the additional middle-class income is an important contribution, which, however, concerns a very limited pool of individuals; this positive income integration can also be seen as having two possible shortcomings: (1) increasing inequality between propertied and property-less middle-class individuals, and/or (2) leading marginal groups to make a living just by renting a room and therefore entirely dropping out of the labour market. A study by Airbnb in London found that over time listings became more common in neighbourhoods that were poorer and less highly educated, although these offerings did not attract many guests (Quattrone et al., 2016). A qualitative empirical study based on fieldwork conducted at four sites (interviews and participant observations at a time bank, a food swap, a maker space and an open-access education site) aimed to analyse how class and other forms of inequality operate within this type of economic arrangement (Schor et al., 2014). The authors find considerable evidence of distinguishing practices and the deployment of cultural capital (i.e. some individuals did not share with others who made grammatical errors in texts exchanged online). This exercise of class power undermines the ability to forge relations of exchange and reduces the volume of trade. This yields inconsistency between actual practice and the widely articulated goals of openness and even equality, which the authors call ‘paradox of openness and distinction’. There is also growing evidence that platforms may enable person-to-person discrimination by race. A statistical analysis of datasets constructed from Airbnb (combining pictures of all New York City landlords on Airbnb with their rental

58    Platform Economics prices and information about quality of rentals) finds what can be seen as indirect evidence of racial discrimination (Edelman & Luca, 2014). The main finding is that, controlling for other relevant covariates, non-Black hosts charge approximately 12% more than Black hosts for quality-equivalent rentals. These effects are robust when controlling for all information visible in the Airbnb marketplace. These findings highlight the existence of discrimination in online marketplaces as an important unintended consequence of a seemingly routine mechanism for building trust. Schor and Attwood-Charles (2017) conclude their recent review of this topic in this way: ‘Every study we have seen that tests for racial discrimination finds evidence of bias’ (p. 9). Another aspect that is highly uncertain and a controversial issue with regard to whether the ‘sharing economy’ has positive redistributive effects in terms of consumers’ welfare for lower social groups. Evidence to support this claim is lacking or inconclusive, but this claim was trumpeted in a an article, entitled ‘Sharing economy benefits lower income groups’, in Financial Times (Bradshaw, 2015); the article cites a modelling simulation as proof that the ‘sharing economy’ will benefit lower income groups and have a democratising effect in terms of access to goods and services (Fraiberger & Sundararajan, 2015). However, as we show later, this model only represents a first exploration that does not warrant the conclusion in the newspaper headlines. Labour relations in the sharing economy are one of the most controversial issue (Hill, 2015b; Ravenelle, 2016; Slee, 2015). According to the lenses of the economists, platforms by increasing labour market efficiency would create more employment opportunities for all (Horton & Zeckhauser, 2016a) and even lead to a society of ‘micro-entrepreneurs’ (Sundararajan, 2016, p. 176). On the other hand, crtics call ‘micro-entrepreneurs’ the ‘new precariat’ (Kuttner, 2013). Investigative journalists’ reports have shed light on the conditions of these ‘on-demand’ workers (Singer, 2014b; Weber & Silverman, 2015b) and showed that in some cases their earnings are not as high as that of companies’ such as Uber and Lyft claim (Weiner, 2015b). In the United States, this issue is currently the object of hundreds of court cases about the misclassification of workers as contractors. The risks are of exploitation and precarity (Scholz, 2016a). Some see the platforms as restructuring of labour relations by shifting risk from firms to workers (Standing, 2011). In Chapter 3, we show unequivocally that the evidence is at best inconclusive and does not support either of the more radical and opposing claims of better opportunities for all and especially less advantaged groups or of full precarisation or exploitation.

Environmental and Socio-economic Impacts In order to introduce the empirical evidence discussed in this section, we present an ex ante theoretical analysis of the possible effects of sharing platforms (effects of digital labour platforms are presented in Chapter 3). Individuals, as providers, can let their assets, sell goods or offer their labour to businesses or to consumers. Conversely, individuals as consumers can rent homes and cars, buy goods, share rides on BlaBlaCar, use Uber instead of traditional

Labour Intermediation Platforms    59 taxis, pay for personal services, and businesses can hire on-demand workers for digital work. Ex ante economists assume that the digital matching between users and providers may deliver a wide variety of efficiencies such as reducing transaction and search costs, improving allocative efficiency, reducing information asymmetries and producing price efficiencies. Considering the consumer markets (i.e. rental, goods and labour-intensive services to consumers), platform-­generated efficiencies may result in consumer welfare effects (increased access at better prices). Furthermore, renting assets, selling goods, driving passengers, performing errands and delivering other personal services are sources of income. Looking at the provision of on-demand labour to businesses, platforms may produce social welfare effects in terms of increased efficiency at aggregate level for both labour markets (better and more matches between supply and demand, mismatches avoidance) and production (unbundling of tasks and lowering of geographical barriers favour vertical and international human capital specialisation, and lower coordination costs favour outsourcing); these may have spill-over effects on employment levels and quality, and on productivity and eventually on growth. There are, however, other more ambiguous effects whose net impact or direction cannot be assumed ex ante and can only be verified empirically. The green and sustainability effects of platforms, for instance, are subject to contrasting forces: reuse of goods and car-sharing in principle reduce emissions, but if improved accessibility produces an aggregate increase in travelling, and the income obtained from platforms is spent on more consumption, then the net sustainability impact could be negative. The net distributive and employment effects are also hard to assume ex ante. Platforms disrupt incumbent industries, and by reducing their revenues may lead to loss of secured jobs that may not be fully offset at aggregate level by the income and flexible jobs created by labour platforms and consumer welfare effects. The net distributive effects of all income-generating activities and consumer welfare can either reduce or increase the level of inequality, depending on how they spread across different social groups. They can have either equalising or polarising effects. For instance, the income integration from renting for middleclass families shown by a study that Airbnb commissioned a former member of the Obama administration to carry out may have equalising effects with respect to upper-middle-class families and polarising ones with regard to propertyless middle-class and lower-middle-class families. Labour platforms may increase or decrease employment levels and quality (i.e. security and social protection) and income polarisation not only for and among on-demand workers but also for regular workers (i.e. competition between platform and regular employment affecting wage levels). The overall impact on labour may be different depending on other effects that ex ante are also ambivalent; for instance, depending on how ‘super star’ or ‘long-tail’ effects play out, they could increase income for a limited group and lower it for others as a result of competition between platforms and traditional employment for most-sought workers; the overall impact on employment also depends on the extent to which platforms favour or not generalise outsourcing and firm’s boundary contraction. This quick broad-brush illustration should have clarified the complexities of potential interactions. The following paragraphs show how limited the empirical

60    Platform Economics evidence is, as it is available only for some of these interactions. Together these two considerations further underscore how simplistic and obfuscating are both the prescriptive narratives and the current public debates, and how further empirical evidence is needed to support policy-making and regulatory initiatives. For instance, some of the grand narratives about the sharing economy and particularly P2P rentals, including those about its contribution to sustainability and consumer welfare, require answers to a number of questions. For instance, how does ‘sharing’ affect ownership and usage of resources? Does it unequivocally decrease ownership levels, decrease usage or both? Under what conditions? Who benefits the most: owners or renters? To what extent would a profit-maximising platform, through its choice of rental prices, improve social welfare? To what extent do frictions, such as moral hazard (additional wear and tear renters place on rented resources) and inconvenience experienced by renters affect platform profit and social welfare? What determines the rental rate and the quantity exchanged in a P2P rental market? How much total surplus is ‘unlocked’ by the P2P rental market, and how is it distributed? When there are substantial bringing-to-market costs (such as labour, excess depreciation and transaction costs), who bears them, and how does it affect short- and long-term equilibria? To these questions, so far, answers come only from theoretical economic modelling. One of these answers is based exclusively on solving equilibrium equations from economic theory (Benjaafar, Kong, Li, & Courcoubetis, 2015), and another two are very partially corroborated by empirical data (Fraiberger & Sundararajan, 2015; Horton & Zeckhauser, 2016b).

Environmental Impacts The environmental benefits of the ‘sharing economy’ are often presented as obvious, and are much advertised in platforms’ own promotional descriptions. In practice, however, the evidence is scant and it is extremely challenging and complex to demonstrate at aggregate level the net impacts in terms of environmental sustainability (Schor, 2014). First-order effects can reasonably be expected to be positive: staying in existing accommodation would reduce the construction of new hotels and/or work spaces, while sharing tools or goods would reduce the production of new goods, both of which should reduce ecological and carbon footprints. Yet, a net impact at aggregate socio-economic level should also consider second-order effects: What happens with the extra money providers earned with the ‘sharing economy’, or users saved? As seen, Airbnb to demonstrate its impacts on city economies provides ‘evidence’ that its guests spend more than traditional tourists, which is self-defeating with respect to the claim that it produces environmental benefits. Only one empirical study of these kinds of impacts was found (Martin & Shaheen, 2010). Some preliminary estimates have been produced for France (Demailly & Novel, 2014), and two modelling simulations have been carried out: one focussed on New York (Santi et al., 2014) and the other on Teheran (Seyedabrishami, Mamdoohi, Barzegar, & Hasanpour, 2012). One study evaluated the greenhouse gas (GHG) emission changes that result from individuals participating in a car-sharing organisation across the United States using data from

Labour Intermediation Platforms    61 a survey and plugging them into an estimation model (Martin & Shaheen, 2010). The authors report a measurable reduction in GHG emissions when a small fraction of households reduces substantially their emissions. This is, however, almost totally offset by the fact that, for the majority of households, car-sharing expands access to cars and obviously increases emissions. According to estimates using data from the French Environment and Energy Management Agency (ADEME), if sharing models in France were operated under the most favourable conditions, savings of up to 7% in the household budget and 20% in terms of waste could be achieved (Demailly & Novel, 2014). A modelling simulation of the effects of carpooling, which calibrates data for New York and makes assumptions about take up and removal of barriers, concludes that in positive scenarios cumulative trip length could be cut by 40% or more, resulting in decreased service cost and emissions (Santi et al., 2014). The modelling simulation of car fuel saving produced by carpooling in Teheran calibrates the data obtained through a stated preference survey (Seyedabrishami et al., 2012) in a model. Given the user preferences revealed and assuming that appropriate strategies to help users identify suitable rideshares would be adopted, the author concludes that carpooling could increase by 30%, which would reduce annual fuel consumption by about 240 million litres.

Socio-economic Impacts In the literature, no empirically robust and comprehensive cost-benefit studies were found which weigh consumer welfare benefits and additional income for suppliers against reduced revenues and jobs for incumbent industries, the cost to the public budget from tax base erosion or future expenditure to provide social protection to on-demand workers. Obviously, at this stage of development in the evidence base, it would be unrealistic to expect this analysis for the entire ‘sharing economy’. However, this kind of aggregate cost-benefit analysis is not available even for a single platform. Neither is there any conclusive evidence on the impact on labour market and on re-distributive effects. While prescriptive and critical essays on these topics abound, only a very few empirical (or modelling) studies are available: (1) As anticipated, as regards consumer welfare, one theoretical economic modelling is exclusively based on solving equilibrium equations from economic theory (Benjaafar et al., 2015), and another two are very partially corroborated by empirical data (Fraiberger & Sundararajan, 2015; Horton & Zeckhauser, 2016b). Of the latter two, only Fraiberger and Sundararajan (2015) consider distributional effects. (2) A quasi-experimental study on the impact of Uber in reducing DUI accidents (Greenwood & Wattal, 2015). (3) A quasi-experimental study of Uber’s competitive pressure effects on traditional taxi industry (Wallsten, 2015). (4) A statistically descriptive analysis of the negative impacts of Uber on taxi industry in three US urban areas (Bond, 2015). (5) A quasi-experimental study of Airbnb impacts on the hotel industry in Austin (Zervas, Proserpio, & Byers, 2014).

62    Platform Economics (6) A quasi-experimental study of Airbnb’s impacts on the hotel industry in Norway, Finland and Sweden (Neeser, 2015). (7) An econometric study (Farronato & Fradkin, 2015) of Airbnb’s two effects on hotel industry: ‘expansion’ (meeting demands of previously under-served consumers) and ‘stealing’ (attracting consumers away from conventional suppliers). (8) One panel study of Airbnb’s effect on tourism industry employment in the US state of Idaho (Fang, Ye, & Law, 2015). On digital labour market effects, a more dedicated analysis is presented in Chapter 3.

Consumer Welfare and Distributional Effects In the theoretical model by Benjaafar et al. (2015), consumers are allegedly always benefitted from the P2P rental. However, this does not necessarily imply sustainability outcomes. According to their model, depending on rental price, P2P rental can result in both higher ownership and higher usage. It is also possible for ownership to decrease, but usage to increase. Only under very specific circumstances, it is possible for both ownership and usage to decrease. Hence, except in the latter case, there is no guarantee that there is economisation in the usage of resources and in the related activities producing emissions. On the other hand, according to this model, it cannot be ruled out that profit-maximising platforms may not have an incentive to completely eliminate moral hazard (by the renters causing wear and tear on the rented assets possibly because of renters’ potential negligence and mishandling). This means that some elements of consumer protection (for the owners) may be circumvented by platforms in order not to decrease volume of transactions (from which they draw their revenues). Horton and Zeckhauser (2016a) present a theoretical model similar to that of Benjaafar et al. (2015), which they partially test with data from a survey on a convenience sample drawn from Amazon Mechanical Turk. These authors model the choice of owning or renting with and without the presence of digital platforms for P2P rental. They also find that ownership and usage can either increase or decrease, depending on various key parameters. Whereas the Benjaafar et al. (2015) model also considers matching (i.e. how utility of being owner or renter depends on the likelihood of finding a match in P2P rental platforms), this is not done in the Horton and Zeckhauser (2015) model. Another difference is that the Horton and Zeckhauser (2015) model looks at how the rental rate influences owners’ and renters’ decisions on how much they use given goods, which is not done in the Benjaafar et al. (2015) model. Despite these differences, the Horton and Zeckhauser (2015) model also predicts that consumer welfare will increase because of the presence of digital P2P rental under all configurations and hypotheses. Their model predicts surplus increases (both in short- and long-term market equilibria for P2P rental) compared with a situation where digital P2P rental was absent. Whereas owners consume less, they are significantly compensated by the income they receive in a way that more than offsets the loss of utility as consumers.

Labour Intermediation Platforms    63 When renters value the goods almost as much as the owners (therefore they demand these goods), having the chance to rent them represents a consumer surplus for them. An important aspect to mention from Horton and Zeckhauser (2015) concerns how BTM costs affect P2P rental markets and platform strategies. For an ideal-typical example of Airbnb, BTMs include the following: cleaning up an apartment, doing check-in and checkout, depreciation from usage plus the fee charged by the platforms (covering conventional transaction costs inherent in finding trading partners, coming to terms and executing payments). BTMs can obviously increase or reduce renters’ supply curve. From the platform’s perspective, lowering these BTMs is convenient when the demand is elastic, since in this case, as the supply curve shifts out, demand increases without reducing price, which increases the revenue from the platform’s fees. The implication is that when demand is elastic, competition between platforms would produce more benefits for consumers and society (i.e. one dominant platform would not lower its fees even when demand is elastic). Last but not least, from both models, it can be deduced that all of these potential gains depend on the possibility of matching. Neither model ensures that matching is always perfect or that all the excess capacity is put to work. Empirical evidence shows that matching frictions and inefficiencies are widespread. These cast serious doubt on the optimistic prediction of both these two theoretical models. It cannot be ruled out, for instance, that ‘super star effects’ prevail over ‘long-tail effects’. In this case, only a small percentage of renters would receive a relatively higher proportion of earnings from renting. In this respect, it is worth recalling that the earlier cited inquiry conducted by the Office of the Attorney General of the State of New York on Airbnb’s operation in the city (based on administrative data forcefully obtained from the platform) found that (a) as little as 6% of Airbnb hosts offer up to hundreds of unique units, getting 36% of private short-term bookings, and receiving $168 million, equal to 37% of all host revenue, in NYC; (b) bookings in just three community districts in Manhattan – the Lower East Side/Chinatown, Chelsea/Hell’s Kitchen and Greenwich Village/ SoHo – accounted for approximately $187 million in revenue to hosts, or more than 40% of private stay revenue to hosts during the review period; (c) by contrast, all the reservations in three boroughs (Queens, Staten Island and the Bronx) brought hosts a revenue of $12 million – less than 3% of the New York City total (Schneiderman, 2014). Fraiberger and Sundararajan (2015) model short-term P2P rental markets and calibrate the model using transaction and survey data from Getaround (one of the leading US P2P car rental platforms), integrated with official US statistics on car ownership, the second-hand car market and patterns of car usage to provide first empirical assessment of the welfare implications of these kinds of markets. They model various effects and costs that can be summarised as follows: As regards positive effects, P2P rental for cars may (a) increase allocative efficiency by creating new gains from trade between consumers; (b) produce surplus for consumers who cannot afford ownership; (c) shift consumption towards higher quality products and (d) lead to surplus for manufacturers by inducing new ‘ownership for peer-to-peer rental supply’. On the other hand, P2P rental could also

64    Platform Economics induce more rapid depreciation and hurt manufacturers because of lower equilibrium of production volumes if durable goods are used more efficiently. There are obvious limitations in the dataset used and some technical aspects of calibration (i.e. estimates of transaction cost function and that of depreciation rate) which may heavily shape the obtained results. Having made these limitations clear, the key results from the modelling simulations are briefly reported below: ⦁⦁ New car ownership would drop (–5%), as would ownership of second-hand

cars (–12%) even under a P2P rental adoption scenario where only 25% of the overall pool of potential users adopt P2P rental. The shift from ownership to rental would be more pronounced among those who use the car below the average usage rate. ⦁⦁ People with above-median incomes tend to maintain car ownership for longer period than people with below-median income. ⦁⦁ The installed base of cars in the economy drops but usage intensity increases, especially of older car. This occurs as the car market becomes more efficient, and, in this respect, it is worth noting that, according to the simulation, belowmedian income car owners significantly increase their supply of cars to the rental market. Consumer surplus increases across the board, but especially for individuals with below-median income as a result, according to the model, of the following: (1) Lower-income consumers who could not afford to own a car and were thus excluded from participation now consume through the P2P rental marketplace. (2) A different fraction of below median-income consumers shifts from being owners to being non-owner renters, realising ownership cost savings, gains from greater usage efficiency and higher-quality consumption. (3) A small fraction of below median-income consumers switches from being non-owners to being owners, induced in part by lower-used car prices, realising surplus gains through their supply activity on the P2P rental marketplace. The simulation also clearly shows that usage will increase and so will emissions of CO2, which once again undermines the claimed sustainability effects. Given the preliminary, exploratory and very partial empirical validation of the model, headlines such as ‘Sharing economy benefits lower income groups’ (Bradshaw, 2015), are exaggerated at best. Airbnb. Zervas et al. (2014) used data obtained for the Austin areas from both Airbnb and the hotel industry. They exploited the significant spatio-temporal variation in the patterns of Airbnb adoption across city-level markets to adopt a counterfactual identification strategy (’Difference-in-Difference’ strategy). They find that Airbnb’s impact on the hotel market is in the order of an 8–10% reduction in revenues. This is non-uniformly distributed, and lower-priced hotels and hotels that do not cater for travel business are the most affected segments. They also find that affected hotels have responded by reducing prices, an impact that benefits all consumers and not just participants in the ‘sharing economy’.

Labour Intermediation Platforms    65 Neeser (2015) replicated the same study as Zervas et al. (2014) with data on Sweden, Norway and Finland. The chapter uses a difference-in-difference strategy with many time periods and different levels of treatment. The data are used to differentiate among Airbnb listings and to identify which type of hotel customers Airbnb is more likely to attract. The main findings are that (a) Airbnb does not significantly affect hotel revenue per available room on average; (b) it contributes to reduction in the average price of a room where Airbnb is mostly present and (c) it is relatively more attractive for foreigners than locals. Farronato and Fradkin (2015) found that the market expansion and business stealing effects of Airbnb differ by location, and attribute this heterogeneity to supply constraints – both legal and geographic – relative to the level of demand. According to the competition model developed by the authors, hotels and P2P suppliers differ in their fixed (higher for hotels) and marginal costs (higher for P2P suppliers). Having run the model, the authors were able to conclude that efficient market structure depends on the level and variability of demand. They were also able to quantify the welfare gains from P2P entry in the accommodation industry. Uber. Exploiting the natural experiment created by the staggered entrance of Uber in different Californian cities between 2009 and 2013, Greenwood and Wattal (2015) adopted a difference-in-difference identification strategy to estimate Uber effects on DUI accidents. They concluded that Uber services reduced alcohol-related motor vehicle homicides. Wallsten (2015) used Google trends as proxies to measure demand for Uber services. They also looked at administrative records of taxi complaints placed by consumers in New York and Chicago for improved service quality by the traditional taxi industry. They identified a negative correlation (increased usage of Uber correlates with fewer complaints) and hazard the conclusion that Uber’s competitive pressure has led traditional taxi drivers to improve their customer service. Bond (2015) analysed Uber impacts in San Francisco, District of Columbia and New York using extensive statistics on the taxi industry in these areas pre- and post-Uber (statistics are used only descriptively and there is no design/attempt to document causal effects). The descriptive data show that Uber has a clearly negative impact on both taxi industry revenues and the value of the medallions.

Ratings and Platforms Functioning: Self-regulation? In Chapter 1, we anticipated a discussion of how reputational ratings work and their limits, which have been recently reviewed (Tadelis, 2016). So, below we just focus and provide more details on specific sharing economy cases. The evidence clearly challenges the claim that sharing platforms do not require consumer protection regulation, thanks to ratings as a form of self-regulation. It further shows that also the claims about the efficiency and market neutral meritocracy of these platforms are greatly exaggerated. Matching and ratings have been analysed recently in an emerging body of micro-economic studies focussing on Airbnb, TaskRabbit and oDesk (Cullen & Farronato, 2015; Einav et al., 2015; Fradkin, 2014; Fradkin et al., 2015; Horton, 2014; Horton & Golden, 2015). The key characteristic of these platforms is the

66    Platform Economics trade-off between minimising transaction costs for users (i.e. search and deliberation) and optimising the use of information for matching the two sides when there is a high level of heterogeneity in what is offered and who demands it (Einav et al., 2015). This trade-off entails designing the platform so that it centralises or decentralises choice. As the authors explain, Airbnb has adopted a decentralised design because in the case of renting apartments it is justified by differences in preferences and in seller costs. Uber, on the other hand, needs to match customers with rides in real time (especially at peak hours). The type of cars and drivers is probably less important than getting a ride at right time, which justifies its centralised design at least with respect to the goal of maximising matches and revenues. Pricing mechanisms can also help in coping with the trade-off between transaction costs and efficient use of information. Indeed, three studies show that Airbnb (Fradkin, 2014), oDesk (Horton, 2014) and TaskRabbit (Cullen & Farronato, 2015) are characterised by (a) high level of heterogeneity; (b) frictions; (c) high percentage of non-matched potential and (d) congestion (i.e. a match falls through because of multiple requests at the same time). Fradkin (2014), for instance, reports the following in Airbnb: ⦁⦁ Potential guests typically view only a subset of potential matches in the market

and more than 40% of listings remain vacant for some dates.

⦁⦁ Hosts reject proposals to transact by potential guests 49% of the time, causing

the potential guests to leave the market, although there are potentially good matches remaining. ⦁⦁ If there were no search frictions (i.e. guests had all market options and knew which hosts were willing to transact with them), there would be 102% more matches, and revenue per searcher would be $117 higher.

In TaskRabbit, Cullen and Farronato (2015) found that the auction mechanism is not very efficient, as it does not vary much with market conditions. They suggested that a simpler mechanism may be preferable; this market clears due to suppliers’ elasticity: in periods when demand doubles, sellers work almost twice as hard, prices hardly increase and the probability of requested tasks being matched falls only slightly. Similar results were found by Horton (2014) for the oDesk market for professional services. These authors seem to suggest that this type of P2P market is inherently frictional but no data and analysis of this kind is available for Uber. Another important mechanism is that of reputational ratings (henceforth simply ratings). These are the evaluative reviews (usually five-star rating systems) that the two sides of a platform make for each other (i.e. in Airbnb, hosts rate guests and guests rate hosts). They are both a source of information on ‘product quality’ and trust. Exchange among strangers is one of the salient characteristics of ‘sharing economy’ platforms, and building trust to get both sides of a market on board has been a key challenge and driver of success for the biggest players such as Airbnb and Uber. As we have seen, motivations to participate are mixed and the ‘sharing economy’ is not unequivocally based on social capital and generalised trust as there is self-interest, and positive and negative reciprocity, and

Labour Intermediation Platforms    67 opportunistic behaviour cannot be ruled out. Therefore, it is obvious that the generalised trust that makes the ‘sharing economy’ possible is the combined result of users’ attitudes and of how these attitudes are effectively leveraged by online reputational rating systems. The reliability of reputational rating systems is a regulatory relevant topic, as it is claimed that they reduce information asymmetry and are a reliable form of self-regulation. It is also claimed that they ensure consumer protection and security and should not be altered by any form of regulatory intervention (Allen & Berg, 2014; Koopman et al., 2014; Thierer et al., 2015). In practice, however, as we anticipated in Chapter 1, there are two main potential biases: under-provision of ratings, and strategic behaviour in providing ratings. Several empirical studies present findings on reputational ratings in sharing platforms (Cullen & Farronato, 2015; Fradkin et al., 2015; Horton & Golden, 2015; Lauterbach et al., 2009; Overgoor et al., 2012; Zervas et al., 2015). The first two studies focussed on CouchSurfing and use big data scraped from the web. They conclude that there is a bias towards positive reviews and that there can be collusive reciprocity among individuals belonging to the same network (Lauterbach et al., 2009; Overgoor et al., 2012). A comparison of the distribution of reviews for the same property on both TripAdvisor and Airbnb shows that ratings in the former are lower than those on Airbnb by an average of at least 0.7 stars (Zervas et al., 2015). More generally, the rate of five-star reviews is 31% on TripAdvisor and 44% on Expedia (Mayzlin et al., 2014) compared to 75% on Airbnb. This difference in ratings could be interpreted as showing that the twosided review system induces bias in ratings. A recent study, involving researchers affiliated with Airbnb, documents through field experiments conducted on Airbnb itself that there is some bias. However, this study also shows that when the bias is removed through experimental treatments, the five-star rating on Airbnb remains substantially higher than even 44% (Fradkin et al., 2015). This would imply that these are a reliable measure of quality to inform other consumers. The study of another platform (oDesk) documents through a laboratory experiment that reputational ratings are fairly inflated (Horton & Golden, 2015). The evidence is thus inconclusive and mixed, and further evidence is needed to know whether reputational ratings are a sufficient and reliable measure of quality and consumer protection, especially in European contexts. Finally, in an empirical analysis of Airbnb’s data and a controlled experiment, it was found that the more trustworthy the host is perceived to be from her photo, the higher the price of the listing and the probability of its being chosen significantly (Ert et al., 2016). The study also finds that a host’s reputation, communicated by her online review scores, has no effect on listing price or likelihood of consumer booking. Further, the authors demonstrate that if review scores are varied experimentally, they affect guests’ decisions, but the role of the host’s photo remains. This last piece of evidence casts serious doubts on the fact that reputational ratings really function as quality assurance and that market rationality prevails in the matching process. The effect of photo may be euphemistically considered as based on heuristics and producing bias if not compared with other mechanisms that are cited in a study earlier lead to race discrimination.

68    Platform Economics

From Legal Battles towards Regulation and Policy? As regulation and policy lagged behind, the unfolding of the sharing drama, courts and judges have supplemented regulators and policy makers. While some regulatory developments are taking place, still sharing platforms remain mostly a topic for litigation. As the majority of litigations concerns labour issues directly or indirectly (even court cases disputing Uber’s claim to be a technology and not a transportation company de facto have labour law implications), these will be considered in more details in Chapter 3. In Europe today, different ‘regulatory’ solutions are emerging in different countries as a patchwork of judges’ decisions and local authorities’ initiatives. In certain European cities, for instance, Airbnb can operate only at the condition of withdrawing due taxes, whereas in other areas strict permissions and licences are required for those who want to put their house on the platform.24 Also, other regulatory matters concerning consumer protection are dealt with a patchwork of different decrees and interpretations of pre-existing laws and regulations in different ways depending on the country.25 The rhetorical polarisation between the ‘passions’ and the ‘interests’ has soon turned into real conflict, legal disputes and consumer protection concerns. As empirically documented (McNeill, 2016), one place where this is being played out most visibly, creating urban policy tensions and conflicts, is the San Francisco Bay area. McNeill (2016) reconstructs the political processes and tensions surrounding the rise of San Francisco as a city of unicorns. He underlines the important role played by technology and venture capital in the political economy of urban development. The urban policy tensions associated with the evolution of new ‘sharing economy’ firms such as Uber and Airbnb, have, according to the author, aggressively challenged municipal regulations in taxi and property rental fields. In addition, legal disputes have given rise to various forms of protests related to negative externalities in neighbourhoods and to shortages and rising prices in the long-term housing rental market. The ‘sharing economy’ giants Uber and Airbnb have been the object of legal challenges and bans. Violent protests by taxi drivers have erupted in many United States and European cities, leading to Uber being banned by local decree or court rulings (Europe). Airbnb has also been challenged for not respecting city regulations in New York, for violating zoning rules and for indirectly contributing to the erosion of the local government tax base. Now, the Airbnb website informs and

24

Three impulse papers requested by European Commission on how Airbnb operates in different European cities provide an exhaustive picture of this fragmentation of approaches (Ranchordás, 2016; Rating Legis, 2016; Smorto, 2016). The most recent overview of regulatory approaches in the tourism sector for the sharing economy could be found in a report published in 2018 by the European Commission (VVA & Spark, 2018). 25 Two reports published by the Commission provide the most updated and exhaustive review of regulatory developments in EU28 concerning consumer protection (Hausemer et al., 2017) and of cases concerning sharing economy platforms that have been dealt by the Court of Justice of the European Union (Psaila et al., 2017).

Labour Intermediation Platforms    69

Box 3:  Conflicts, Bans and Court Cases. ⦁⦁ Taxi companies complaints against Uber Technologies have led to the

⦁⦁

⦁⦁ ⦁⦁

⦁⦁ ⦁⦁

⦁⦁

prohibition of Uber in several cities, including Berlin and Brussels, since April 2014 (Vasagar, 2014). In Germany, the Frankfurt District Court went further than previous cases on the topic, and highlighted the fact that Uber did not have necessary licences and insurance. It was therefore competing unfairly with the local taxi industry (Scot & Eddy, 2014). Uber is banned in Spain, and risks being banned in the Netherlands (Kroet, 2014). It is also a very sensitive issue in Brussels (Keating, 2014). On 26 May 2015, a judge in Milan blocked UberPop in Italy on the grounds that it represents ‘unfair competition’ (http://milano.corriere. it/notizie/cronaca/15_maggio_26/uber-pop-bloccati-app-servizio-tuttaitalia-vittoria-­tassisti-277dd376-038b-11e5-8669-0b66ef644b3b.shtml, accessed on June 8, 2015. Uber is operating in London, but the clash between the interests of licensed taxi drivers and Uber operation is far from resolved (Collins, 2014). In New York City, Airbnb has faced several problems. Its hosts have been fined (Liber, 2013) and a very strong and negative report on its practices in the city was released by the NYC state attorney office (Schneiderman, 2014). The report concludes that according to New York regulations, Airbnb rooms and apartments qualify as illegal hotels. It also contains the following information: (a) As little as 6% of Airbnb dominated the platform, offering hundreds of unique units, receiving 36% of private short-term bookings and earning $168 million (37% of the host’s revenue; (b) Airbnb Rentals displaced long-term housing for about 5,000 apartments; (c) bookings in just three community districts in Manhattan – Lower East Side/Chinatown, Chelsea/Hell’s Kitchen and Greenwich Village/SoHo – accounted for approximately $187 million in revenue to hosts, or more than 40% of private stay revenue to hosts during the review period. By contrast, all the reservations in three boroughs (Queens, Staten Island and the Bronx) brought hosts a revenue of $12 million – less than 3% of the New York City total. So far, in Europe, Airbnb seems to have been hit less hard than Uber, although it has been seriously constrained in Barcelona. In Amsterdam, however, it has been regularised (Ranchordas, 2015, p. 8).

requires hosts to be aware of local laws and their landlord’s rental policies, and comply with them, which may prohibit short-term rentals (Miller, 2016; Zrenner, 2015). Furthermore, Airbnb has also started to collect taxes in some US cities and in Amsterdam. The most important cases in Europe, however, have concerned Uber with both national level decision and one important recent decision by the Court of Justice of the European Union.

70    Platform Economics Various cases that would be too lengthy to be illustrated here have raised regulatory concerns; for instance, liability and insurance, identification, licencing and permits, safety standards, reliability of reputational ratings, information and privacy, frauds, etc. These aspects are discussed in more depth in terms of regulatory approaches and open issues; therefore, these are selectively and briefly mentioned below. Incidents reported for Uber drivers and/or with Airbnb hosts have raised concerns about the fact that suppliers of lift and rental services are not required to obtain licences, permits or certification (Rauch & Schleicher, 2015; Sablik, 2014). It is not clear, for instance, whether the platform is liable when a hired car crashes or a host’s apartment is damaged (McLean, 2015), or whether it is responsible for the security of the service provided to users by the platform. Furthermore, platforms, in order to escape liability, can argue that they are only intermediaries providing a ‘matching service’, and are not direct service providers (Malhotra & Van Alstyne, 2014). Ratings can be biased and inflated and it is possible that platforms present the results of search in a way that is more convenient to them than to the users (Einav et al., 2015). With respect to the utilisation of data by the platforms, Einav et al. (2015) pose the following questions: Can consumers limit platforms’ use of data? Can platforms share/sell ratings and purchase history? What about potential gender and race discrimination in ratings, leading to these groups getting fewer opportunities? It is obvious that small NFP platforms with a few thousand members and platforms worth billion dollars, engaging millions of users and providers, are very different and pose different regulatory and policy challenges. However, there are also stark differences between commercial platforms between, for example, providing ride services (i.e. Uber: who is liable? and what type of insurance could cover the drivers?) and giving a ride in your own car for a fee covering the costs (i.e. BlaBlaCar: the car owner is liable but fully covered by standard insurance). On the one hand, proponents of self-regulation argue that formal regulation is costly and serves to protect vested interests. On the other hand, the proponents of extending the reach of formal regulation to P2P platforms argue that this would correct market failures that private parties on their own cannot overcome. However, more moderate approaches are also evolving. Consensus is growing around the idea that the ‘sharing economy’ cannot be left entirely unchecked, nor can it be regulated by means of traditional command and control approaches. Aside from this more general debate, there are still several unresolved issues that are briefly summarised below (except for labour issues that are addressed in Chapter 3): (1) Taxation. Substantive law for tax-sharing activities exists, but enforcement may present challenges because (a) some platforms opportunistically pick the most favourable regulatory regime and (b) micro-providers raise unique compliance concerns. Airbnb is currently engaged with legislators in drafting or adjusting existing legislation. In addition, its website requires hosts to be aware of and comply with local laws and their landlord’s rental policies, both of which may prohibit short-term rentals (Miller, 2015; Zrenner, 2015). Furthermore, Airbnb has also started to collect taxes in some US cities and in Amsterdam.

Labour Intermediation Platforms    71 (2) Negative externalities, liability and insurance. Negative externalities for ride service platforms are derived from unsafe and uninsured or under-insured driver/car. Short-term accommodation rentals produce negative externalities for neighbourhoods (increased traffic, parking places occupied, noise, tenants disturbing neighbours, etc.) and by removing properties from long-term rental markets. Liability and insurance, however, are not only matters of negative externality but may also concern the two sides of a ‘sharing’ transaction. The issue is again to determine who is liable if something goes wrong, and to guarantee that ‘sharing’ activities are insured. It is reasonable to expect that some intervention may be needed to define liability, ensure safety and close the insurance gap. Under specific circumstances, the negative externalities of short-term rentals should also be addressed. (3) Information asymmetries and cognitive biases. Various information asymmetries, exacerbated by typical cognitive biases documented in the behavioural economic literature, cast doubt on the extent to which self-regulation can fully protect consumers. This entails various more specific issues such as the reliability of reputational ratings, safety standards, frauds, dispute resolution and redresses. The chances are that consumers will make poor decisions when faced with an overwhelming range of choices, poor regulation and unclear avenues for recourse in case of a dispute. They may also fail to fully appreciate risks and safety requirements. Under these circumstances, regulation and/or nudges could help increase protection for consumers. (4) Licencing and certification schemes. While licencing and certification schemes tend to be ineffective and may unduly favour incumbents, serious incidents with both Uber and Airbnb have caused critics to demand that they be imposed on large commercial platforms. Platforms have tried to boost confidence with ID checks and vetting processes. There are doubts, however, as to how transparent and rigorous these inspections are. (5) Data and privacy. There are concerns about the amount of data that ‘sharing’ platforms are collecting from consumers, given the sensitive nature of some of these data and how these are being used. (6) The potential implications of competition law. From the evidence reviewed on the characteristics and functioning of the largest platforms, it seems that market dominance is out of reach for most of them due to heterogeneity and matching frictions. It is not so unlikely, however, for Uber. On the other hand, improvements in the matching algorithms, together with pricing strategies and use of personal data without any regulatory checks, may change the situation and make market dominance more likely for a few other platforms.

This page intentionally left blank

Chapter 3

Digital Labour Markets in a Broader Perspective Introduction Whereas in Chapter 2 we took self-defining practice at face value and engaged with them in an operation of deconstruction and attempted reconceptualisation, in this chapter we refrain from using expressions such as ‘crowdsourcing’, ‘crowd employment’ and other value-loaded terms. Here, we refer to market operators for what they really are, namely digital labour markets where labour-intensive services are traded by matching requesters (employers and/or consumers) and providers. We even prefer the term ‘market’ to ‘platform’ as the latter is surrounded by a politically motivated rhetoric (Gillespie, 2010) matched by that on the objectivity of the algorithms they use (Gillespie, 2014). Furthermore, as we anticipated in Chapters 1 and 2 and discuss at the end of this chapter, given the control exerted by some of these operator on the exchange of labour taking place, there are doubts that they could be considered two-sided markets (2SMs) and tend to resemble in some respects vertically integrated firms or input resellers. Using neutral terms and a crisp typology also helps placing these operators in the correct current and historical perspective. From the 1990s until the start of the Great Recession in 2007–2008 in OECD (excluding the United States for which data are not available, and including EU21), non-standard work (NSW) as a whole (part-time work, temporary work and self-employment)1 accounted 1

Although this is not the main topic of the book, our analysis in this chapter is formed by relevant streams of literature on NSW, including, among others, institutional reports (Eichhorst et al., 2013; Eurofound, 2015; European Commission, 2013, 2014a, 2014b, 2015c, 2016a; European Trade Union Institute (ETUI), 2012, 2015; ILO, 1997, 2015; OECD, 2010, 2014, 2015a, 2015b, 2016), monographic works (Emmenegger, Palier, & Seeleib-Kaiser, 2012; Esping-Andersen, 1990; Esping-Andersen & Regini, 2000; Sapir, 2005; Standing, 2011; Thelen, 2014) and various theoretical and empirical scientific articles (Allmendinger, Hipp, & Stuth, 2013; Baranowska & Gebel, 2010; Brewster, Mayne, & Tregaskis, 1997; Buschoff & Protsch, 2008; de Graaf-Zijl, van den Berg, & Heyma, 2011; DiPrete, Goux, Maurin, & Quesnel-Vallee, 2006; Eichhorst,

Platform Economics: Rhetoric and Reality in the “Sharing Economy” Digital Activism and Society, 73–122 Copyright © 2019 by Emerald Publishing Limited All rights of reproduction in any form reserved doi:10.1108/978-1-78743-809-520181004

74    Platform Economics for about 50% of all job creation and 60% extending from the crisis year until 2013. On average, 33% of total employment in OECD countries is in the form of NSW with wide ranging differences among countries: as low as 20% in Eastern Europe to up to 46% in the Netherlands (OECD, 2015a, p. 137). So, the kind of work practices emerging from the activity of digital labour markets cannot be considered as a separate ‘silos’ totally unrelated to the diffusion of NSW and the associated trend towards ‘casualization of work and demutualisation of risks’ (De Stefano, 2016, p. 6); as observed in one OECD (2016, p. 22) chapter on these digital labour markets, most of the work mediated by such markets ‘is likely to be some form of NSW’. First, we define and focus on digital labour markets: (1) that work as digital marketplaces for non-standard and contingent work; (2) where services of various nature are produced using preponderantly the labour factor (as opposed to selling goods or renting property or a car); (3) where labour (i.e. the produced services) is exchanged for money; (4) where the matching is digitally mediated and administered, although performance and delivery of labour can be electronically transmitted, or physically and (5) where the allocation of labour and money is determined by a collection of buyers and sellers operating within a price system2. Four types are identified based on two key dimensions: (a) whether tasks 2013; Ferber & Waldfogel, 1998; Giesecke, 2009; Green & Livanos, 2015; Hipp, Bernhardt, & Allmendinger, 2015; Ichino, Mealli, & Nannicini, 2008; Janine, 2009; Kahn, 2012; Kalleberg, 2009; Kautonen et al., 2010; Nunez & Livanos, 2014; Smeaton, 2003; Van Lancker, 2012). There are various other expressions used to refer to what we will always indicate as NSW, such as, for instance, non-traditional employment relations (Ferber & Waldfogel, 1998), flexible working practices (Brewster et al., 1997), atypical or non-standard employment (Buschoff & Protsch, 2008), NSW (e.g. OECD, 2015, p. 138), non-standard forms of employment (Eurofound, 2015; International Labour Office (ILO), 2015); not surprisingly, there is no universally accepted definition of NSW arrangements (OECD, 2015a, p. 138). Broadly defined, these include all employment arrangements not conforming to ‘traditional’ and ‘normal’ ones that are full-time, regular, open-ended employment with a single employer (as opposed to multiple employers as in the case of work obtained through temporary agencies, or, most recently, through online labour platforms) over a long time span. Operationally, Labour Force Survey statistics include and report data on NSW as comprising the following broad types: (a) on account self-employed (i.e. self-employed not hiring employees); (b) temporary or fixed-term contracts and (c) part-time work. Hence, pragmatically, we use NSW as the main expression and define it operationally following Eurostat practice. 2 This definition excludes various online players. Online matching for traditional jobs, as performed by LinkedIn, is excluded by condition (1), since the matching of people is for highly contingent work. Condition (2) excludes services delivered using a decisive physical capital or goods component, such as Airbnb (renting a room or a house), RelayRides (renting a car) or eBay and Etsy (selling second-hand and bespoke goods) but includes Uber and Lyft as they are at the centre of labour-related disputes and because labour is preponderant over physical asset. Various forms of online self-defined crowdworking without traditional monetary remuneration are also excluded by condition (3). Disguised crowdsourcing, contest-based crowdsourcing or markets, expert networks or platforms where professionals collaborate in production processes as peers and charitable and public interest crowdsourcing (i.e. public-funded

Digital Labour Markets in a Broader Perspective    75 are entirely traded and delivered online, or are traded, monitored and paid online, but delivery is physical and collocated and (b) what type of tasks are traded and what skills are required to deliver them (low skills, mostly routine or manual vs high skills and mostly cognitive and interactive). On the basis of the first dimension, we distinguish online labour markets (OLMs), which are potentially global, from mobile labour markets (MLMs), which by definition are localised. Using this distinction together with that concerning the type of task traded, we obtain four types: (1) OLMs for micro-tasking (i.e. small pieces of routine cognitive work requiring low to middle levels of skills as traded, for instance, in Amazon’s Mechanical Turk (MTurk); (2) OLMs for tasking and at times delivery of entire and self-contained projects (i.e. tasks requiring middle to high skill levels such as in Upwork or Freelancers); (3) MLMs for physical services (i.e. performing low-skilled manual work and errands such as in TaskRabbit) and (4) MLMs for interactive services (i.e. interactive services requiring high skills such as in TakeLessons). Second, this chapter reviews both rhetorics and some hypotheses that can be derived, especially from the economic literature. As these markets concern labour issues, they have been the source of most of the legal disputes, and the rhetorics are very much entangled with some of the hypotheses of economic and social science literature. These rhetorics and hypotheses include, among others, the idea that individuals work in these markets for pin-money or out of other non-monetary motivations, that these are stepping stones towards better employment or entrepreneurship, that they bring back to work the unemployed or under-employed (including students, stay-at-home parents, retirees, etc.) and that they bring about a flat world online meritocracy, thus making labour markets in general more efficient. On the opposite and more pessimistic sides, critics see these markets as new forms of precarisation, exploitation and deprivation of any labour-related rights and social protection. Third and fourth, the above rhetorics and hypotheses are contrasted first against some descriptive quantitative findings (where we present data from a novel research on NSW and social protection from Codagnone et al., 2018b) and then against some experimental and quasi-experimental studies. The available empirical evidence is still limited and so it enables to debunk some of the rhetorics and ex ante hypotheses, whereas on others it shows that more

platforms for innovators) are excluded. Disguised crowdsourcing refers to the fact that most Internet users, without knowing, perform small tasks when browsing, buying online or playing games that, behind the veil of software, are used commercially (Cherry, 2011; Felstiner, 2011). In contest-based online markets such as ‘99Designs’, ‘InnoCentive’ and ‘Article One Partners’, the buyers propose a contest for specific objects (i.e. solutions to engineering problems in InnoCentive) and select the winner among the solutions proposed by different ‘workers’ and pay for it. Condition (4), on the other hand, extends a definition previously formulated only for OLMs to also include what we call MLMs. It extends the definition proposed by Horton (2010, p. 516) to include labour markets such as Uber/Lyft and TaskRabbit/Gigwalk. The last condition (5) typically applies to a 2SM.

76    Platform Economics empirical research is needed before making sweeping generalisation in one sense or another. Providers of labour-based services in these markets tend to be younger and better educated, but there is no predominance whatsoever of students and retirees. Overall, women are either more or equally represented than men, although gender stereotypes penalise them even in anonymous contexts due to heuristics and biases reinforced by the underlying algorithms, which evidently refute the hypotheses of the advent of a flat world online meritocracy. In terms of the employment status of providers, it emerges from recent research a polarisation between self-employed and individual with open-ended full-time employment; whereas the latter fact is somehow unexpected, this polarisation suggests that these digital markets are not bringing unemployed or underemployed back to work. They do not work for ‘pin money’ or out of boredom, as the available evidence shows that there are fairly large shares of individuals for whom earnings from working for digital labour markets represent their primary source of income and/or who engage in a portfolio of several activities. On the other hand, also considering the motivations, the more pessimistic views that working in these digital markets is widely an involuntary choice are not fully supported by the available evidence. Fifth, we consider the current and the future trends of these forms of work intermediated by digital labour markets in the broader context of NSW, considering also the evidence available in terms of potential positive and negative effects of these forms of employment and how it relates to those exchanges in digital labour markets. Finally, as in Chapter 2, we conclude considering legal disputes and emerging regulatory and policy debates and actions with respect to both the status of labour providers (contractors vs employees) and the issue of access to social protection.

Conceptualisation and Dimensional Relevance The commercial operators that we discuss in this chapter go under various labels not only in media but also in academic articles and policy documents. Such labels include, among others, the ‘sharing economy’, the ‘collaborative economy’, ‘crowdworking’, ‘crowdsourcing’, the ‘gig economy’ and the ‘on-demand economy’. On the other hand, in the economic literature, the platforms that allow the full process from start to delivery to be completed digitally (i.e. Amazon’s MTurk (AMT), Upwork and Twago) are defined as two-sided OLMs that are potentially global (Horton, 2010). These platforms that are driven digitally but require a final physical delivery of services (i.e. TaskRabbit, Helpling, Uber, etc.) have also been defined and treated in economic literature as two-sided local labour markets but have been given different names: ‘mobile crowdsourcing markets’ (Musthag & Ganesan, 2013), ‘mobile crowdsourcing marketplaces’ (ThebaultSpieker, Terveen, & Hecht, 2015) and ‘on-demand mobile workforce’ (Teodoro et al., 2014). We opted for two neutral names: OLMs and MLMs. Using the two dimensions anticipated in Introduction chapter, we propose a conceptual heuristic tool, which is simple and parsimonious but grounded empirically and theoretically (see Fig. 5). The four types are as follows:

Digital Labour Markets in a Broader Perspective    77 OLM

Mturk

Twago

Lower skills

Higher skills

Helpling

Takelessons

MLM

Fig. 5:  Heuristic Typology. Source: Authors’ Elaboration. (1) OLM Micro-tasking (i.e. AMT, but also crowdflower and many others) (2) OLM Tasking (i.e. Twago, but also Upwork, Freelancers etc.) (3) MLM Manual/Interactive Services (i.e., TaskRabbit, Helping etc.) (4) MLM Cognitive/Interactive Services (i.e. Takelessons) The usual acknowledgement of the limits inherent to any conceptualisation of an emergent and little studied phenomenon is probably pleonastic. The empirical manifestations of the digital labour markets are certainly more nuanced than what can be captured by any typology, considering also that fast and continuous evolution makes such manifestations a moving target. There are more fine grained distinctions and sectorial categorisation (Vakharia & Lease, 2013; Valenduc & Vendramin, 2016) that for our interest are not necessary, and actually may prevent seeing the most important aspects of what we are analysing. Considering the two dimensions of our typology, we briefly argue why they are the salient ones for our interest. The above matrix uses two key dimensions. Whether the work can be entirely performed online, or eventually requires physical delivery, is important on two counts. First, from a regulatory perspective of global OLMs where employers, platforms and workers could be based in different countries may prove to be more difficult to regulate and would require possibly some international governance framework. On the other hand, mobile and local work, given that the production, the providers, and the consumers are all located physically in one country, may be more easily addressed by existing legislation and there are higher chances that crowdworkers would manage to organise themselves and have a representation.

78    Platform Economics Furthermore, the type of occupational safety and health risks tend to be different. In the case of fully online work, matters of ergonomics and problems with long hours in front of a computer are a major risk. In the case of manual labour work delivered physically, workers are exposed to other risks such as physical injuries while building or assembling something; moreover, as some of these local services involve consumers, there are also issues of insurance and liability. Finally, leaving aside the rarer case of local services requiring higher cognitive skills (i.e. take lessons), local mobile crowd employment concerns manual labour and can also be interpreted in terms of the Routine Biased Technical Change (RBTC) hypothesis (Autor, 2008, 2013; Autor & Dorn, 2013; Autor, Katz, & Kearney, 2006; Autor, Levy, & Murnane, 2003). The second dimension defined in terms of skills involved is a proxy used to simplify the matrix but clearly embeds consideration concerning the routine or non-routine nature of work as well as the extent to which a piece of work is broken down into micro- and macro-tasks, or left as a self-contained project. Also, this dimension has clear regulatory relevance, because very high skilled contractors may be at ease with working as freelancers, whereas crowdworkers involved in low level routine cognitive tasks, or in manual work, may be in need of more social protection. The main prediction is the emergence of a U-shaped job polarisation with relative decline in routine cognitive and manual jobs and increase of high-level non-routine cognitive and non-routine manual jobs. This aspect is not unrelated to the emergence and the future of digital labour markets, as we have discussed later when we placed them in a wider context and perspective. OLM micro-tasking. Electronically transmittable cognitive micro-tasks paid per piece are traded in markets such as MTurk, Clickworker, Crowdflower, etc. Typical work pieces include object classification, tagging, transcriptions, marketing spam, data entry, content review, editing, website feedback etc. Micro-tasks are highly standardised repetitive tasks, and require low- to middlelevel skills. In these markets, small pieces of work are put out in high volumes, with correspondingly low compensation levels. The individuals involved as providers could be workers misclassified as contractors. In this respect, it is worth noting that Crowdflower recently Crowdflower had to pay $500,000 in compensation to some contractors for violation of minimum wage legislation in the United States.

Box 4: Online Micro-tasking Ideal-typical Functioning. Stylised using MTurk.3 Requesters post small tasks that cannot be computerised. Humans find it easy to tell if two different descriptions correspond 3

As noted Vakharia and Lease (2013), researchers remain fascinated with Amazon’s MTurk and most academic papers focus on this platform. While MTurk opened the field in 2006, currently a myriad of other platforms offers alternatives. Nonetheless, MTurk is described here as a typical case. MTurk is a platform where ‘requesters’

Digital Labour Markets in a Broader Perspective    79

to the same product, tag an image with descriptions of its content or transcribe a high-quality audio snippet. However, these tasks, usually called HITs, are still hard for computers. Employers post tasks, specify the range of data for processing, define the structure of the form into which the data must be put, create a set of instructions for workers and assign the tasks at a price. Computers can use a programmable Application Programme Interface (API) to post tasks on marketplace, which are then fulfilled by human users. This API-based interaction gives the impression that the task can be carried out automatically. Workers find and perform tasks on the platform, which sends the output directly to employers’ IT systems without human intermediation. The employer defines the criteria that candidate workers must meet to access the task, and if these criteria are met, workers do the job without the need for a digital application process. These criteria include the worker’s approval rating (the percentage of tasks the worker has performed, employers have approved, and consequently paid for the same), the worker’s self-reported country, and whether the worker has completed certain skill-specific qualification exams offered on the platform. This filter approach to choose workers allows employers to request work from thousands of temporary workers in a matter of hours. Once a worker submits completed work, the employer can choose whether to pay for it or not. Workers have no access to employers’ data, nor they know for what larger purpose the micro-tasks are completed.

(employers) can post micro-tasks (Human Intelligence Tasks, or HIT) such as object classification, tagging, transcriptions, marketing spam, data entry, content review, editing, website feedback and much more. Individuals performing these tasks are called ‘turkers’. The 500,000 registered turkers make on average of up to $5 per hour. The participation agreement, which both requesters and turkers must sign, is the only governing agreement (although all participants must have an Amazon account) and stipulates that turkers complete tasks as independent contractors. All juridical rights are placed with the requesters. Amazon can cancel an account any time for violation of the terms of participation agreement and the worker may be deprived of any remaining earnings. Amazon declines all responsibilities related to the transactions between requesters and workers in terms of quality, safety or payment issues, and explicitly states the following: ‘You use the site at your own risk’. MTurk maintains an ‘acceptance rate’ for each worker so that requesters can recruit workers with higher rates of task acceptance from prior requests. Once the worker’s bid for a given HIT has been accepted, it must be completed within a defined timeframe. However, there is no time limit in which firms should evaluate the task or provide reimbursement. The ‘mandatory satisfaction’ clause gives the requester the authority to reject an HIT without any justification and without payment. At the same time, they can access the work without forfeiting ownership. The requesters typically include the academic community (using turkers as subjects of surveys and experiments), start-ups and entrepreneurs, large corporations often using intermediary (consulting) agencies aggregating tasks and controlling quality.

80    Platform Economics OLM tasking. Electronically transmittable tasks (and in some cases fully selfcontained projects) paid with fixed contract per deliverable (more often) or per hour (less often) are traded in markets such as Upwork and Freelancers. Typical requested work includes software development, engineering and data science, graphic design and clerical and secretarial work. Some tasks require middle-level skills, and are fairly routine tasks, while others demand flexibility, creativity, generalised problem-solving and complex communications (i.e. high-level skills). While there is no data to substantiate it, it is reasonable to assume that some contractors are truly highly skilled freelancers, whereas others are not too dissimilar from those working for the markets of quadrants 1, and could possibly be considered as misclassified workers. And yet, so far Upwork and similar outlets have not been the object of any legal dispute.

Box 5: Online Tasking, Ideal-typical Functioning. Stylised using Upwork.4 Registered contractors provide their CVs and profile pictures, which are stored digitally and gradually integrated with their work 4

Upwork and freelancers are the ideal typical examples of online labour market for more complex tasks and sometimes for working on entire projects. Both Elance and Guru were launched in 1999, followed by oDesk in 2005 and Freelancer in 2009. In 2014, Elance and oDesk merged into Elance-oDesk, which today has become Upwork. While there are some platform-specific features, these online labour markets share several characteristics. They allow employers to hire short-term workers by registering with the platforms and posting jobs. Registered contract workers then bid for the posted jobs and advertise their skills and experience on profiles pages. The platforms maintain a rating system, and track records of the work completed. They earn revenues by charging a transaction fee (percentage of transaction) or membership fees (most often only from contract workers; in some cases from both employers and workers). oDesk is further described here as a typical case. For registered employers, oDesk has information on company name, legal representative or owner, location and industry. Employers are free to post as many jobs as they need and are required to specify task description, employer’s location and the type of contract offered. oDesk supports both hourly wages (where employers must specify the expected number of hours per week and number of weeks needed for completion) and fixed price (where they must specify budget and deadline). In case of hourly wages, oDesk offers strict monitoring (up to keystrokes) enabled by its virtual office software. Registered contractors provide their names, contact details and set up profile pages detailing their skills, education, work experience outside oDesk, oDeskadministered test scores, certifications, whether they belong to an agency and oDeskspecific work histories and feedback scores. They can apply for jobs by submitting cover letters and bids to job postings. Employers can interview and negotiate bids

Digital Labour Markets in a Broader Perspective    81

history on the platform: offered bids, hours worked and feedback ratings from previous engagements. More disaggregated information is also available, such as per-contract feedback, hours worked and earnings. Employers create job posts with titles and descriptions, nature of work, the skills required (from a menu of existing categories) and, where foreseen, contractual form (lump sum or hourly wages). Once the post is sent, the platform reviews it and posts it online. Furthermore, the platform makes both job posts and contractors searchable by various categories to help the search on both sides. For employers, verified attributes (number of past jobs, average wage rate paid, etc.) are also made available to the contractors. The process of matching can work in two ways. Contractors search for jobs and apply for the one they like, after which they appear to the employers in, what is usually called, an ‘applicant tracking system’. Employers then look at all the information on the applicants. Conversely, employers can search the contractor database and try to recruit the desired profile directly. The search functionality usually allows employers to search by skills and other attributes and it returns lists of contractors with relevant information about their work history. Employers can then send recruiting invitations to the selected contractor(s). It is possible that direct recruitment by employers tends to concentrate on highly positive selected contractors (these have more experience, higher past wages, greater earnings, etc., and consequently accept higher hourly wages than contractors who autonomously apply for a job). Recruited contractors have no obligation to accept a job, as employers have no obligation to hire a contractor that has applied autonomously. Once match occurs, the platform intermediates the relationship in various ways, including through tracking software that contractors install in their computers and functions as a digital punch clock, allowing hours worked and earnings to be measured essentially without error. with applicants before hiring and can hire as many contractors as they like. Once hired, the contractor completes tasks remotely. Contractors submit their work online to employers and are paid via oDesk. Employers have the option to give contractors bonuses and can reimburse expenses through oDesk. The platform charges $8.75 per transaction. oDesk contract workers are highly educated and come from a large number of countries across the world. Those from lower-level income countries are hired mostly; 90% of the employers requesting work on oDesk, according to a 2013 survey, were SMEs. The number of employers billing on the platform per quarter increased by over 800% between 2009 and 2013, and the number of working contractors per quarter increased by approximately 1,000% over the same period. The quarterly wage bill on oDesk increased by approximately 900%: from $10,000,000 to almost $100,000,000 over the same period. The average hourly wage in software development ($16) was approximately double than that of writing and translation ($8) and more than triple than that of administrative support ($4), customer support ($5) and sales and marketing ($5).

82    Platform Economics MLM physical services. Tasks requiring physical delivery of mostly manual services demanding low- to medium-level of skills and paid with fixed contract per task (more often) or per hour (less often) are traded in markets such as TaskRabbit, Gigwalk, etc. (several others launched and operating in Europe). Various misclassification lawsuits (contractors vs workers) have affected these MLMs in the United States. A particular case in this quadrant is represented by rides services such as those provided by Uber or Lyft that have been the centre of misclassification lawsuits. Box 6: MLM Physical Services’ Ideal-typical Functioning. Stylised using TaskRabbit.5 Buyers post description of a domestic task they intend to outsource and sellers can search through posted task by location 5

A description of how TaskRabbit worked between 2009 and mid-2014 is provided here. Currently, the platform is present in 20 US cities and in London with a pool of about 30,000 individuals performing tasks as independent contractors and 1.2 million requesters. In the language of the platform, ‘posters’ outsource tasks to ‘rabbits’, who search the posted offers on city-specific lists. Until the change was introduced in 2014, posters could request any sort of task, even the oddest kind. The five largest categories between 2009 and mid-2014 were shopping and delivery (24%), moving help (12%), cleaning (9%), home repairs (6%) and furniture assembly (4%). It must be noted that TaskRabbit also allows tasks that can be delivered digitally and do not require direct interaction, such as editing texts, carrying out usability testing of mobile apps, etc. (during the period considered, these tasks made up 10.4% of the total). In the original model, matching took two forms: a task was posted and then the posters accepted the first offer, or posters asked for bids as in an auction model. Posters and rabbits went through a vetting process, although the screening was more rigid for the latter. For the former, the platform checked the identity through social networks and their payment method and capacity. The latter were subjected to a background check, they answered a digital survey (on motivations, skills and availability) and were interviewed by TaskRabbit employees to assess their fit. The acceptance rate of applications by potential rabbits was on average 13.6% (although it varied greatly from month to month). This screening was, however, reduced in the Spring of 2013 in an attempt to involve more rabbits. On average, 78% of posted tasks received an offer (on average 2.8 offers); of these, 63% were successfully completed at an average price of $57. TaskRabbit charged (from rabbits) a 20% commission fee of successful tasks. The unit of observation is city–month (a poster is active if he/she posts at least one task in a given month for a given city, and in the same way, a rabbit is active if he/she submits at least one offer within that city–month). On average, there are 708 active posters and 255 active rabbits in a city–month. Typical posters are predominantly females (55% of buyers) and relatively affluent. The modal poster is a woman aged between 35 and 44 years with a household income between $150,000 and $175,000. The rabbits are younger and not surprisingly have lower incomes. The modal rabbit is 25–34 years old and has a household income between $50,000 and $75,000. Finally, it should be noted that TaskRabbit recently introduced a wage floor and an insurance policy. It is now not possible to earn less than $12.80 an hour (this is higher than any possible minimum wage in the United States). A new insurance scheme covers both sides for possible property damage or bodily harm up to $1 million.

Digital Labour Markets in a Broader Perspective    83

and respond with an offer. Buyers post any possible kind of task (shopping and delivery, moving help, cleaning, home repairs, furniture assembly, etc.), which require local and short-notice delivery. The matching process works in one of the two ways: the buyer posts a task price and accepts the first offer, or asks for bids and selects from the prices offered by the sellers. Given that buyers and sellers must meet, there is a vetting process, which is not needed in online markets. Identity and payment methods are checked, with the screening process being more rigorous for sellers. Past history on the platforms and ratings are available for both sellers and buyers. MLM interactive services. This could, in principle, be an empty set for local digital markets for highly skilled services requiring complex communications, so far limited to localised matching between students and teachers providing lessons in person as in the case of Takelessons. For both types (1) and (2), the overwhelming majority of transactions are peer-to-business, own their own account, are self-employed, and are requesting tasks also as a business entity. On the other hand, the majority of transactions in type (3) are ‘peer-to-peer’ in the sense that requests come from individuals as consumers; in this respect, however, it must be noted that some of the MLMs in this quadrant do work requiring mobility (mystery shopping, inspection services, etc.) for businesses. Although digital labour markets can be defined and categorised in much clearer way as compared to the other ‘sharing platforms’ discussed in Chapter 2, it is equally difficult to come up with robust and firm numbers about how many people find works through these and an overall market value of transactions and the revenues accruing to these commercial operators. There are no official statistics and administrative data, given the novelty of the phenomenon and the grey area in which some of these markets operate. One way to proceed to estimate the dimension of the phenomenon in terms of people involved as providers of labour-intensive services is to identify a large number of the most important platforms and then extract figures of registered providers from each. To do this, the first step is the identification of all relevant digital labour markets, which is not an easy task. Evans and Gawer (2016) estimate that globally there are 300 operational ‘workplace platforms’. This small number makes sense considering that the winner-takes-all dynamic ‘digital economy’ leads market structure with a limited number of platforms with some noteworthy dimension (Degryse, 2016; Kuek et al., 2015). Furthermore, due to the recent creation of these markets, they tend to be quite unstable: merging, disappearing and becoming inactive, and so one is ‘shooting at a moving target’: 80% of operators surveyed in a recent study on EU28 were created after 2010 (Fabo et al., 2017). The first estimate presented by Codagnone et al. (2016a, pp. 21–23) placed the total number of registered providers globally at about 52.6 million at the end of 2015, and noted that digital markets with large numbers of contractors were mostly originated in the United States, although the largest one (i.e. Upwork) also

84    Platform Economics has registered contractors from European (EU) countries. De Groen et al. (2017) estimate that there may be 12.8 million people active in ‘online work-related platforms’ in EU28, although they admit that this number may be inflated due to over-reporting by platforms and/or to double counting providers who are registered with more than one platform (De Groen et al., 2017, p. 351). Kuek et al. (2015, p. 19) have estimated that only 10% of registered users should be expected to be active workers, which if applied to the above two estimates would reduce them, respectively, to 5.6 million globally and 1.3 million in EU28. Some interesting data on digital labour markets gathered by a recent Joint Research Centre study identifying about 200 platforms in EU28 (Fabo et al., 2017) are as follows: ⦁⦁ Altogether 169 (84.5%) platforms were of EU origin and others originated in

the United States but are operational in Europe and tend to be the largest ones.

⦁⦁ In France and the UK 50 markets were identified; in Germany, the Netherlands

and Spain about 40 platforms; in Belgium and Italy 30; whereas in all other countries the number was 20 or less. ⦁⦁ In 35% of cases, the work assignment is performed by the markets, and some of them exert a tight control over both providers and requesters. ⦁⦁ The most common business models are based on percentage of commission fee per service (40%) or flat fee per service (36%). The alternative way to attempt a sizing exercise is through surveys, which, however, often use different samples and channels (Computer-Assisted Telephone Interviewing vs Online) and different definitions, and pose very different questions; as a result, the numbers vary widely and do not replicate when comparing different surveys conducted in the same country. As observed, ‘the variation of the findings is considerable, ranging, for example, from 0.4% to 19% of the population (Florisson & Mandl, 2018, p. 12). Below we selectively report the main findings of a few of these surveys for illustrative purposes only: ⦁⦁ A McKinsey (2016) online survey covering Germany, France, Spain, Sweden,

the UK and the United States led to an extrapolation of 9 million people who earn money on these digital markets between EU15 and the United States. ⦁⦁ A multi-country online survey reports the following country-wise prevalence and associated extrapolation:

–  Austria: 9% use digital labour markets weekly (540,000) –  Germany: 6% (3.5 million) –  Italy: 12% (5.3 million) –  The Netherlands: 5% (600,000) –  Sweden: 5% (310,000) –  Switzerland: 10% (600,000) –  The UK: 5% (2.2 million)

⦁⦁ The survey designed on NSW forms at the end of 2017 by one of the authors of

this book and mentioned in the Introduction (Codagnone et al., 2018b) found that 25% of those surveyed and are not unemployed in 10 European countries

Digital Labour Markets in a Broader Perspective    85 (France, Germany, Italy, the Netherlands, Poland, Portugal, Romania, Slovakia, Spain and Sweden) make income from sharing platforms, of which 15% sell their work and 13% sell goods or renting space (multiple answers, 3% do both). Prevalence for these countries in common with a previous survey are: Germany 13%; Italy 13%, the Netherlands 7.8% and Sweden 8.7%. So, there is a large variation (in excess) for Germany, whereas prevalence figures are similar for the other three countries. ⦁⦁ The Chartered Institute for Personnel and Development (CIPD, 2017) surveyed 5,019 people between the age of 18 and 70 years across the UK in traditional employment, platform work or unemployed. The survey found that 4% of all respondents had found work through digital labour markets at least once over the past 12 months, which if extrapolated would amount to 1.3 million respondents. In the same vein, as per the similar part in Chapter 2, we conclude that digital labour markets are certainly not marginalised statistically and can be captured in surveys, but we are still far from having a reliable quantification.

Rhetorics and Economic Hypotheses The operators intermediating labour-intensive services have aroused, if possible, even more rhetorical controversies than other ‘sharing platforms’, and certainly many more legal disputes, as we show in the final section of this chapter. As they concern labour, they have become a point of sharp controversies (Hill, 2015b; Ravenelle, 2016; Slee, 2015). As sources of inspiration for the different rhetorics, on the one hand one finds the hypothesis of market economic efficiency (Horton & Zeckhauser, 2016a) and the growth of micro-entrepreneurship (Sundararajan, 2016), while on the other hand the more sociological perspective on the risks and negative effects of exploitation and precarity (Scholz, 2016b). From the review of media accounts (newspapers, magazines, blogs) performed, it emerges how the enthusiasts see these digital markets as empowering millions of individuals to unlock the value of their time, especially for those segments of human capital that are away from institutionalised employment (i.e. ‘flexers’ such as stay-at-home parents, retirees, students, etc.), including the underemployed, unemployed and independent highly skilled professionals. Digital labour platforms are portrayed as helping individuals earn good extra money, and in many cases avoid boredom, achieve work–life balance through flexible and personally chosen work schedules and by working from home, be creative and autonomous, and enabling firms to deal with work picks without incurring unnecessary fixed costs and reach talents not available domestically. The rhetoric of a flat world is integrated with that of the advent of a global online meritocracy, for instance, the Elance-oDesk (2014) (now renamed Upwork) 2014 annual impact report. The pessimists see digital markets as a new unregulated channel of exploitation and saving of labour costs by employers. They argue that the ‘gigs’ traded on these markets are the components of formerly full-time jobs, parcelled up and put out to tender on piece-by-piece basis to increase outsourcing across the board

86    Platform Economics (i.e. of both core and non-core tasks) and reduce labour costs (Felstiner, 2011; Smith & Leberstein, 2015). Others see these markets as creating a new class of networked precariat with no benefits and social protection, contributing to the steady erosion of ‘labour contract’ and to increasing inequality (Berg, 2016; Cherry, 2011, 2016; Kuttner, 2013; Summers & Balls, 2015, p. 32). It is observed how in OLMs work is carried out in a regulatory vacuum, as the global nature of the platforms neutralises national labour laws and there is no international-level agreement (Beerepoot & Lambregts, 2015, p. 246). Actually, as candidly stated by both Horton (2010, p. 517) and Agrawal, Horton, Lacetera, and Lyons (2013a, p. 19), platforms perform government-like functions. Despite the fact that MLMs are localised, they are not regulated yet, although they have been at the centre of lawsuits in the United States (see Cherry, 2016). Hence, on the positive and supportive side, we find four rhetorical discourses. First, the rhetoric of a flat world allowing digital labour migration with no boundaries and a world of online meritocracy. Second, the discourse on extra money as a motivation to work for the flexers (students, retirees, stay-at-home parents, etc.). As acutely observed by Berg (2016, p. 18), the claim about individuals working in digital labour markets for ‘pin money’ or out of boredom is a replication of the rhetoric used in the late 1950s and in the 1960s when a new temporary agency industry in the United States was portrayed as employing just middle-class wives killing time and earning extra money. This is an emblematic case of Hirschman’s (1977, 1982) claim that rhetorical discourses of the past tend to resurface. Third, the alleged contribution of online labour platforms to bring back to work the unemployed and the underemployed. Fourth, discourse on the flexibility, autonomy and creativity that these platforms allegedly provide to their workers. However, on the negative side we have: (a) digital labour markets assigning work that is even less guaranteed than in traditional NSW; (b) more inequality and exploitation; (c) unregulated functioning that undermines the broadly defined labour contracts and (d) several negative social consequences. Following the well-known ‘flat world’ narrative about technology and globalisation (Friedman, 2005), OLMs are presented (Horton, 2010, p. 521) and discussed empirically (Beerepoot & Lambregts, 2015) as drivers accelerating the flattening of labour markets through virtual migration and associated global online meritocracy, and as potentially increasing wage convergence with its potentially positive consequences for workers in developing countries. They are also seen as potentially affecting the international division of labour and deepening human capital specialisation at global level. Whether the world is flat in general, and whether OLMs are really favouring convergence and abating all geographic barriers, is an empirical matter that has to be resolved. It is, however, worth noting that limits to routinisation and digital job-matching were presented much before the emergence of OLMs (Autor, 2001, 2008; Autor et al., 2003). Autor (2001, 2008) expressed scepticism about digital job-matching, as he considered that the structure of these digital labour markets would be ripped with information asymmetry and would not be capable of conveying the ‘high bandwidth’ kind of information needed for a job match. As we show later, some empirical evidence confirms his hypothesis. More generally, the market efficiency view

Digital Labour Markets in a Broader Perspective    87 typical of economics consider several first-order effects that digital labour markets allegedly have: (a) the pool of employers and workers increases as geographical and time differences are abated and (b) search, coordination and transaction costs are reduced as a result of search algorithm and intermediated administration and tracking system. Consequently, the economic narrative goes, more and better matches between employers and workers occur (more because of the efficiency of matching, and better because availability of searchable information could reduce mismatches). Next, as coordination costs decrease, the unbundling of jobs into tasks and tasks into micro-tasks multiply the number of possible ‘gigs’ beyond the number of employers. As a result, more and better matches, lack of distance barriers and unbundling of tasks, human capital specialisation at global level and greater vertical functional specialisation (division of labour, or Smith effects) could be achieved with positive overspills on labour productivity. Another side effect, through reduced coordination and monetary costs, is allegedly the increased possibility of outsourcing (especially for SMEs). All of the above should produce generalised net welfare effects in the form of more efficient labour markets, including increased employment and improved productivity. The more optimistic version of this story is the digital search would benefit not only the unemployed searching for a job but also the inactive or underutilised population (part-time workers who would prefer to work for full time, recently retired people, stay-at-home parents, discouraged workers). Additional positive effects on productivity might originate from more stringent monitoring of workers – especially unskilled – and better incentives for both unskilled and skilled workers.

Descriptive Socio-economic Picture In this section, we provide a socio-economic picture using descriptive evidence from reviewed sources, zooming from new primary data produced by the earlier mentioned survey on NSW, and access to social protection, which focussed on individuals working through digital labour platforms (Codagnone et al., 2018b). At the end, we briefly contrast these descriptive findings with some of the rhetorical discourses illustrated above.

Socio-demographic Contours of Labour Providers Beyond slight differences depending on specific OLMs and MLMs considered, workers tend to be younger and highly educated than their corresponding populations of reference. This type of generalisation cannot be made on gender balance, as it differs substantially depending on the OLM or MLM considered and by country. OLM micro-tasking. The first survey done in 2009, providing information on individuals working for MTurk (henceforth also turkers), found that 50% were from the United States, 40% from India and 10% from other countries (Ipeirotis, 2010). The same survey indicates that the US respondents were fairly represented the US Internet population. They were, however, younger (54% in the 21–35 age group vs 22% in the general Internet population) and mainly females (70% vs. 50%).

88    Platform Economics The US turkers showed lower income (65% with household incomes of less than $60K vs. 45% in the general population) and lived in smaller families (55% with no children vs. 40% with children). The gender split for Indian respondents was reversed (the majority were males). For both the United States and India, turkers’ self-reported educational level was higher than for the corresponding general Internet population. In terms of income, turkers based in India had significantly lower incomes (55% declared an income of less than $10,000/year) compared to those based in the United States. A survey of workers in MTurk (only Indians and Americans) and in Crowdflower (including 50 or more nationalities) broadly confirms the above-mentioned socio-demographic profile (Berg, 2016). This survey finds gender balance among turkers based in the United States but with more males compared to previous surveys (52% males, 48% females), whereas in India, turkers are 69% males; however, Crowdflower’ workers are on average 73% males. With respect to age and education, the earlier profile of turkers is confirmed; in Crowdflower only 14.1% have a high school diploma or less (1.1% have education less than a high school diploma), and most workers have at least attended some years of college (28.4%) or have a college (36.7%) or a postgraduate degree (16.9%). An online survey conducted for the European Parliament (2017b) of providers in Mturk, Clickworker, CrowdFlower and Microworkers found that about 60% of workers were aged less than 40 years. OLM tasking. Apart from a generic statement about workers registered with (at the time) oDesk (now Upwork) being highly educated and skilled (Agrawal et al., 2013a), there are no other scientific source on them, and the only available information comes from reports based on surveys conducted by Elance-oDesk (2014) and Nubelo (2014), a digital labour market for the Spanish-speaking world where clients and contractors mostly come from Spain (accounting for 40.6% of contractors and 64.5% of employers) and to a much lesser degree from Argentina, Colombia and Mexico. The data for Elance-oDesk come from two surveys of its registered contractors conducted in 2014 in nine countries (for Europe, including the UK and Ireland).6 In terms of age, 90% of contractors are below the age of 45 years (26% 18–25; 48% 26–35 and 16% 36–45 years) with only 10% above the age of 46 years (6% 46–55 and 4% aged 55 years and above). Only 23% have less than college education, 49% have a college degree and 28% having a graduate degree; 55% have been working as independent contractors for more than five years and 20% for less than three years; for 63%, working with this digital market represents the primary source of income (complete income for 18%, most of the income for 25% and more than half of the income for 20%) and only 37% indicated that it accounts for less than half of their income. The contractors of Nubelo show similar age profiles as seen above (57% reported as ‘Generation Y’, 27% reported as ‘Generation X’ and 15% reported as ‘baby boomers’), are 6

The report illustration of the survey methodology is succinct and not fully clear to say the least (see Elance-oDesk, 2014, pp. 81–82). The sample was extracted from nine countries representing developed and non-developed countries according to the World Bank categorisation.

Digital Labour Markets in a Broader Perspective    89 mostly males (65% males vs 35% females) and in 60% of cases are college graduates. On average (considering responses from all countries), 61% of respondents work for full time for Nubelo, which is for them the only source of income, 25% have another full-time job and 14% have another part-time job; the average tenure with Nubelo is two years. Although the younger contractors are the most numerous numbers, the analysis of the data shows a correlation between age and full-time work in Nubelo: contractors aged above 45 years work for full time in 72% of cases, whereas this occurs in only 45% of cases for those below the age of 25 years. When considering only Spain, the percentage of those working for full time and obtaining their entire income increases to 74% (vs 61% in complete sample), and the average tenure is also higher. MLMs physical services. In TaskRabbit, the customers on the site are predominantly females (55%) and relatively affluent (Cullen & Farronato, 2015, p. 7). The modal customer is a woman aged 35–44 years, while the providers are younger, typically in the age group of 25–34 years. A comparison of MTurk demographics with those reported in a study of Mobile Crowd7 (Mushtag & Ganesan, 2013) shows that in the latter, men outnumber women (71% men and 29% women, whereas in MTurk, women account for 60%) and, in general, the educational level is higher (75% with college degree vs 55% in MTurk). The ‘taskers’ or ‘agents’ (to use the same expressions as in the chapter) on this platform are young (70% are aged under 35 years). The earlier cited multi-country survey (Huws, Spencer, Syrdal, & Holts, 2017) confirms that young people are overrepresented among labour providers in digital labour markets with percentages above 50% of the sample, although people aged 55 years or more are not marginal (11–17%). In the same survey, female participation is between 40% and 52%, depending on the country. For the UK survey by CIPD (2017, p. 53), young people were 69%, with 44% females. In another UK survey, there were 54% males and 46% females, compared with a general population of 49% men and 51% women (Department for Business, Energy and Industrial Strategy (BEIS, 2018)). Finally, other surveys confirm that providers of labour on these digital markets are more likely than average to have degree-level qualifications (BEIS, 2018; European Parliament, 2017a, 2017b). This holds more for online work compared with locally delivered services, where the educational attainment of participants is closer to that of the average population (European Parliament, 2017a). On the other hand, CIPD (2017) found a small difference in the levels of educational attainment between UK gig workers and the general population. In a survey for the European Parliament (2017b), it was found that more than 50% of the sample comprised individuals who held a third-level degree (both first and second stages of tertiary education), and 23.8% were currently in pursuit of a degree. Interestingly, Kuek et al. (2015, p. 3) found a difference between online tasks through Elance (now Upwork)

7

‘Mobile work’ is a pseudonym the authors used to maintain the anonymity of the platforms. However, from the description of the tasks performed, the author believes that the data come from Gigwalk.

90    Platform Economics workers and that of AMT and CrowdFlower workers: 75% of online professionals had a university degree compared with 33% of micro-task workers – a finding that fully corroborates the typology we proposed.

Motivations Most sources providing robust empirical data on motivations converge in that money is by far the primary extrinsic reason why individuals to work on these digital labour markets, regardless of which specific case one considers. Other extrinsic motives such as flexibility, autonomy and working from home are detected, but are given less importance than money. Intrinsic motives are much less important and the discourse about working to kill time, for fun or for network purposes is mere rhetoric. Even in MTurk, where earnings are very low, various studies show that money is the primary motive (Kaufmann, Schulze, & Veit, 2011; Pilz & Gewald, 2013; Ross et al., 2010); this is further confirmed by longitudinal ethnomethodological studies of online communities such as ‘Turker Nation’ and ‘Turkopticon’ (Irani & Silberman, 2013; Martin, Hanrahan, O’Neill, & Gupta, 2014; Silberman & Irani, 2016), although other aspects, besides money, also emerge, but as side benefits. The most recent survey of MTurk and Crowdflower confirms the importance of money as a primary motive, but find somewhat more support for other reasons such as working from home. Whether the choice of working fragmented gigs as voluntary or involuntary is an issue on which more empirical evidence is needed (but see infra data from Codagnone et al., 2018b). The survey on MTurk and Crowdflower shows that 90% of the sample considers insufficient work as a main concern and would like to work more (Berg, 2016). Lack of steady flow of work is also among the main drawbacks cited by 26% of contractors working for Nubelo (2014) and by 49% of American on-demand workers surveyed in another industry study (RFS, 2015). Other surveys by associations of Freelancers in the United States indicate that working independently (including in digital labour markets) is a choice in 60% and a necessity in 40% of cases (Freelancers Union & Elance-oDesk, 2014; Freelancers Union & Upwork, 2015; MBO Partners, 2015). A number of experiments focussing on MTurk provide some indirectly relevant insights on motivation, perception and actual behaviour. For instance, in natural experiment settings, it is found that motivation depends not only on money but also on whether task monotony is offset by disclosing why these tasks are requested. This is interpreted as the importance of telling workers the meaning of the work (Chandler & Kapelner, 2013). Another experiment found that turkers typically did more work when paid more but did not deliver better results (Mason & Watts, 2010). Two experiments show that when turkers had to think about the responses of their peers, combined with financial incentives, they provided higher-quality results (Shaw, Horton, & Chen, 2011). On the other hand, an earlier experiment had found that workers can attempt to game the system for monetary rewards. Younger men (aged under 25 years) were more likely to engage in gaming, while men over 30 and women of any age were more likely to take tasks seriously (Downs, Holbrook, Sheng, & Cranor, 2010). Another experiment with MTurk shows that workers work less when the pay is lower, but they do

Digital Labour Markets in a Broader Perspective    91 not work less when the task is more time-consuming (Horton & Chilton, 2010). Finally, a simple experiment with turkers explored their perception of employers to challenge the many critical articles about exploitation of turkers (Horton, 2011). The findings reported are that, on average, turkers perceive employers in MTurk to be slightly fairer and more honest than offline employers. However, limitations of this experiment hardly warrant sweeping conclusions.8

Employment Status The earlier cited survey of MTurk and Crowdflower finds that (a) students are only 14.5% of respondents, which contradicts the myth about these markets employing mostly students and other ‘flexers’; (b) 33% of respondents of the sample at the time of the survey (conducted in November–December 2015) were unemployed and (c) 37% reported that working for such digital markets is their primary source of income (Berg, 2016). Other studies, however, report different results and the apparently counter-intuitive findings that many providers of labour in these markets are individuals with full-time open-ended employment status. Huws et al. (2017) report that more than 50% in the seven countries surveyed were in full-time employment besides platform work, except for Italy (41%) and the Netherlands (48%). In the CIPD (2017) survey in the UK, 58% of respondents were permanent employees. A qualitative study of 102 earners on six platforms in the United States found that 26% are dependent on the platform for their primary source of income, 43% are partially dependent and 32% treat the income as supplemental (Schor & Attwood-Charles, 2017). Still in the United States, a national sample survey reports that 29% depended on this income to meet their basic needs whereas 42% said that it was ‘nice to have, but I could live comfortably without it’ (Smith, 2016). This study also found that 44% of gig workers have full-time jobs. On the other hand, all these sources indicate an increasing number of individuals with two or more ‘jobs’. In the UK, for instance, as many as 61% of those who work show a portfolio of activities strategy and are registered with two to five digital labour markets, and 7% are registered with more than five platforms (Huws & Joyce, 2016b).

Working Conditions: Earnings and Other Aspects Earnings. The evidence on the earnings of contractors in these digitally mediated labour markets is limited to a few studies of a limited number of cases (MTurk, Crowdflower, Uber, Upwork, Nubelo and TaskRabbit, some of which

8

A total of 192 subjects (convenient sample) were randomly allocated to answer one of two questions. One group answered a question about the fairness of other employers they had had; the other answered a question about the fairness of employers in MTurk. There is an inherent limit to the reliability of these results due to the sample size and recruitment method. In addition, the results may suffer from ‘experimenter effects’: subjects may have been encouraged to exaggerate how honest and fair they find AMT employers. Furthermore, the subjects were turkers with a 95% completion rating (i.e. the best workers).

92    Platform Economics are self-reported by the operators). In addition, the data from replies to the survey cannot be taken at face value. This evidence, however, can be supplemented with the reported investigative journalistic accounts. According to the first analysis of earning in MTurks in 2010, 10% of the posted micro-tasks in MTurk were priced at 2 cents or less, 50% above 10 cents and only 15% of HITs above $1 (Ipeirotis, 2010). Using a stochastic simulation, turkers’ potential average hourly wages were estimated at $5 (which is lower than the US minimum wages of $7.25 per hour). More recently, it has been confirmed that reported average hourly wages for both MTurk and Crowdflower is between $1 and $5.5, although 10% of turkers in both the United States and India report hourly earnings of above $10 (Berg, 2016). According to Hall and Krueger (2015), Uber drivers earn $6 per hour more than drivers of traditional cabs ($19 per hour vs. $13 per hour). This figure, however, has been seriously challenged by investigative journalism, which, taking into account idle times and running costs, estimates a net earning per hour barely above the minimum wages. In Upwork, the average hourly wages are $16 in software, $8 for writing and translation, $4 for administrative support and $5 for both customer support and sales and marketing. With TaskRabbit, the average job is worth $55 and now it cannot entail an hourly wages of less than $12.50. Journalistic reports show that contractors can make up to $25 per hour (or between $2,000 and $3,000 per month) by working from different platforms (doing errands for TaskRabbit and driving for Uber) as long as they work up to 12–15 hours per day. Considering that there are superstar effects (i.e. 20% of contractors do 80% of the jobs) on many of these digital markets, it is reasonable to assume that, for the majority, average earnings are limited. As shown, the earlier cited surveys conducted in the UK and Sweden confirm based on nationally representative samples that earnings are very modest. In TaskRabbit, the customers on the site are predominantly females (55%) and relatively affluent (Cullen & Farronato, 2015, p. 7). The modal customer is a woman aged 35–44 years, with a household income between $150,000 and $175,000. The providers are younger and, not surprisingly, have lower incomes. The modal provider is aged 25–34 years old and has a household income between $50,000 and $75,000.

Box 7: Working Conditions (Investigative Journalistic Accounts). •  M  icro-tasking platforms have been described as digital machines that turn workers into ghosts (Marvit, 2014), or as horrific digital sweat shops (Uddin, 2012; Zittrain, 2009). Furthermore, research has shown that 90% of tasks posted on AMT are priced at less than 10 cents and, on average, people only make $4.8 per hour (Ipeirotis, 2010). •  Various investigative journalistic reports have shown that providing generic personal and home services through sharing economy platforms (i.e. TaskRabbit) provides no flexibility or work–life balance. Workers have to work for more than 12 hours a day in order to cobble together a

Digital Labour Markets in a Broader Perspective    93

decent income by running errands and driving people around (Kantor, 2014; Shontell, 2011; Singer, 2014a; Weber & Silverman, 2015a; Zimmermann, 2015). •  With respect to the above, a survey of on-demand workers (including those working in the sharing economy platforms) has found that there is no flexibility and autonomy because working hours are demanddependent. Many workers are dissatisfied with both the wages and the work schedules (reported in Smith & Leberstein, 2015, p. 6). •  Investigative work on Uber and Lyft drivers uncovered that (a) many are full-time taxi drivers with their own cars and there is little autonomy and flexibility (Rapkin, 2014); (b) contrary to claims by Hall and Krueger (2015) that Uber drivers make about $16 per hour, field work (including working as an undercover driver) places the net earnings at $7.20 per hour (Brown, 2015; CEPR, 2015; Griswold, 2014; Guendelsberger, 2015; Weiner, 2015a) and (c) documents from court cases on Uber and Lyft unequivocally prove that the advertised autonomy and flexibility is a myth, since the two platforms can terminate drivers if their dispatch acceptance rate is too low. These platforms also look for accounts to deactivate when there are too many drivers or business is slow.

According to the data reported in Berg (2016), the respondents from MTurk and Crowdflower, for whom this kind of work is a primary source of income, lack any form of social security coverage, only 8.1% of those based in the United States report making regular payments into private pensions and only 9.4% contribute to social security. In the 1099 Economy Workforce Report (RFS, 2015), the respondents (all self-employed9) indicate as the most desired benefits in the order of importance: health insurance; retirement benefits; paid sick, holiday and vacation days. The same report shows that 8% of drivers and 16% of delivery workers are uninsured; 30% have no health insurance; 43% complain about insufficient pay and 49% report about not finding enough work. An indirect way to imagine what are social protection conditions of workers in digital labour markets is to look at the situation in traditional NSW where lack of any social protection and benefits, such as unemployment benefits, eligibility for work injury benefits as well as sickness and maternity benefits, is the norm (OECD, 2015b, 179–190). In the United Kingdom, 81% of those who work are main breadwinners and generally make a modest annual income: it is estimated from replies that 42% of those working earn less than £20,000 before taxes, and only 7% make more than £20,000 before taxes, the remaining 51% are spread between these two figures (Huws & Joyce, 2016b). As in the UK, yearly earnings before taxes are fairly modest: 53% earn less than 300,000 KR, 87% earn under 500,000 KR and only 4% make more than 700,000 KR (Huws & Joyce, 2016a). 9

The ‘1099’ label refers to the kind of tax forms that these individuals file.

94    Platform Economics Working for MTurk. A group of broadly defined ethnographic studies provides evidence on what it means to work for MTurk (Bergvall-Kåreborn & Howcroft, 2014; Irani, 2015; Irani & Silberman, 2013; Martin et al., 2014; Silberman, Irani, & Ross, 2010a; Silberman, Ross, Irani, & Tomlinson, 2010b). Taken collectively, regardless of the method and evidence gathered10, they identify the following problems faced by MTurk workers concerning asymmetries in terms of both information and levels for action: ⦁⦁ Information asymmetry in general. MTurk terms of agreement stipulate that the

judicial rights over the task accomplished by the workers pass to the requesters, who can accept or do not accept the output. The platform tracks and maintains workers’ acceptance rates so that requesters can recruit workers who have higher rates of task acceptance from prior requests; however, there is no equal mechanism for workers to filter employers. Workers only see the name the requester chooses to use and receive only limited information about the tasks whereas firms can access the employment history of workers. ⦁⦁ Employers’ moral hazard: When workers submit work to employers through MTurk, they have no guarantee of receiving payment for their work. Employers can retain the work and may not pay without providing any justification. ⦁⦁ Moral valence: MTurk workers have to learn to identify illegitimate tasks to stay safe online. The lack of transparency raises ethical questions as workers are unable to make judgements about the moral valence of their work (Zittrain, 2008). ⦁⦁ Costs of requester and administrator errors are often borne by workers: When a requester posts a task with inadequate instructions, they often do not get the responses they want from workers and reject the work. The responsibility for the lack of quality does not belong to the workers but with either the requester or the platform administrator. Workers discuss these issues in online forums such as ‘Turker Nation’ (Martin et al., 2014) and also on the activist platform ‘Turkopticon’, where they help each other by going public and evaluating their experiences with the employers (Irani & Silberman, 2013). Aside from the empirical documentation of these problems, authors (Bergvall-Kåreborn & Howcroft, 2014; Irani, 2015) have drawn more sweeping conclusions about MTurk. Bergvall-Kåreborn and Howcroft (2014) conclude that MTurk is not merely a passive broker but also an active organiser in a long supply chain. MTurk plays a fundamental role in establishing the conditions for crowdlabour because it makes it possible to exercise control by bypassing traditional routes and regulatory procedures when procuring labour supply. In these authors’ view, this is further corroborated by the fact that a range 10

Some are digital ethnographic analyses of online forums such as ‘Turker Nation’ (Martin et al., 2014) or ‘Turkopticon’ (Irani & Silberman, 2013). Others are based on both online ethnography and e MTurk websites and industry sources (BergvallKåreborn & Howcroft, 2014). The more in-depth and sustained analysis presented by Irani (2015a; 2015b), on the other hand, is based on four years of involvement in crowdsourcing by the author as both participant and observer.

Digital Labour Markets in a Broader Perspective    95 of intermediaries who filter work requests from their clients to the MTurk platform have emerged and are very actively organising work on the platforms. Irani (2015) argues that MTurk and other similar platforms are a source of social differentiation within the universe of knowledgeable workers, i.e. between ‘innovative workers’ and ‘menial workers’. She claims that the former maintain their identities as creative, highly valued entrepreneurs by outsourcing tedium, tinkering with labour and casting their work as high-tech entrepreneurs. In addition, she also considers that MTurk is a controlling and organising platform, which makes cheap labour invisible. In other words, Irani (2015) sees MTurk as a platform that helps ‘ameliorate the contradictions of intensified labour hierarchies by obscuring workers behind code and spreadsheets’. Online crowdlabour markets often address issues of risk and mistrust between employers and employees from employers’ perspective, but less often from employees’ position. Based on 437 comments posted by crowdworkers (turkers) on the AMT participation agreement, we identified work rejection as a major risk experienced by turkers. Unfair rejections can result from poorly designed tasks, unclear instructions, technical errors and malicious requesters. Automated workers’ control. Ethnographic (Lee, Kusbit, Metsky, & Dabbish, 2015; Rosenblat & Stark, 2015) and quantitative (Chen, Mislove, & Wilson, 2015) analyses have focussed on control and surveillance by algorithm in Uber. Rosenblat and Stark (2015) argue that Uber’s digitally and algorithmically mediated system of flexible employment builds new forms of surveillance and control, which result in asymmetries around information and power for the drivers. Their analysis casts doubt over the claims made in the Hall and Kruger (2015) that the main reason drivers join Uber is the flexibility of schedules. Their findings are corroborated by Lee et al. (2015) who interviewed drivers from both Uber and Lyft and triangulated these interviews with passenger interviews. They describe how drivers respond to algorithm-assigned work and how they share with each other (on digital forums) informational support and social tactics on how to resist or game the rigid restrictions imposed by the algorithmic control. In addition, a quantitative study of how the Uber surge price algorithm works shows that this algorithm is opaque and clearly manipulated, does not reliably reflect the real situation of peak demand, and is resisted with various tactics by both drivers and passengers (Chen et al., 2015). Based on their analysis, Chen et al. (2015, p. 13) conclude: Our observations about Uber’s surge price algorithm raise important questions about the fairness and transparency of this system. The forces at play on markets such as eBay and Airbnb are well understood: the supply of goods is transparent, and prices are set by competing individuals. In contrast, Uber does not provide data about supply and demand, and the pricing algorithm is opaque. These contributions, taken collectively, document the functioning of an ensemble of surveillance instruments which substitute direct managerial control and create power asymmetries between the platform and the drivers. The pillars of this system are assignment algorithms, surge price algorithms and semi-automated

96    Platform Economics evaluation (i.e. drivers’ acceptance rate plus the ratings received by the passengers). These match three aspects typical of human resources management: work allocation (i.e. passenger assignment plus predictive scheduling), information (dynamic surge pricing) and evaluation (semi-automated evaluation). Passenger assignment, for instance, seems to severely limit both the flexibility and autonomy of drivers for two reasons (Lee et al., 2015; Rosenblat & Stark, 2015). First, drivers are forced into blind acceptance of passengers, since when they accept a call, they are not shown the destination or how much they can earn on the fare. Second, in principle, drivers can refuse a call but they risk being suspended or removed from the system (in both Uber and Lyft). Uber in San Francisco (SF) requires drivers to have cancellation rates below 5% and an acceptance rate of at least 90%. Uber uses predictive scheduling trying to influence drivers (i.e. there is high demand now in the area where you are located) to convince drivers to keep working when they attempt to log off (Rosenblat & Stark, 2015, pp. 8–10). Chen et al. (2015), analysing four weeks of Uber’s surge pricing data in downtown SF and mid-town Manhattan,11 conclude that it seems prone to manipulations, has a few bugs and raises issues of fairness and transparency. These quantitative findings are corroborated by triangulating them with qualitative interviews (Lee et al., 2015) and analysis of online forum posts (Rosenblat & Stark, 2015). It seems that some drivers are not influenced by surge pricing, others avoid surge areas or try to game the system by colluding with passengers. Still others consider it to be an unfair system because at times it leads them into areas where they expect higher earnings, which later do not materialise. According to some interviews, surge pricing changes too rapidly and unexpectedly for drivers to use the information strategically to boost their incomes (Lee, 2015, p. 1607). Though in-depth analysis of management by algorithm as the primary focus was found only in studies of Uber, issues of control, surveillance and standardisation also affect other platforms such as Upwork (Agrawal et al., 2013a) and MTurk (Ipeirotis & Horton, 2011). In platforms, such as Upwork, digital on-demand workers can be controlled even by measuring their productivity in terms of keystrokes (Horton & Tambe, 2015, p. 131). Some platforms include virtual office applications, which ensure tight control of contractors (i.e. with regular screen shots and activity logs). In some cases, contractors are incentivised to log onto these applications by a guarantee of a certain hourly wages (Agrawal et al., 2013a, p. 11). Standardisation and control are presented as key ways of helping MTurk and other similar platforms to scale up by increasing efficiency and reducing frictions (Ipeirotis & Horton, 2011). The practices described above fit in what has been defined as ‘algocracy’, to be a new form of algorithm-based governance alternative for both markets and hierarchies (Aneesh, 2009). Algorithms can be seen as a new source of rhetoric that promises ‘objectivity’ (Gillespie, 2014). Both in the practice of administering matching and quality control processes and in its public relations campaign, Uber

11

They made 43 copies of the Uber smart phone app and distributed them throughout downtown SF and midtown Manhattan to individuals who were acting as drivers and/or passengers.

Digital Labour Markets in a Broader Perspective    97 smartly appeals to the algorithm in both rationalistic and ‘affective’ ways, thus blurring analysis and prediction.

New Evidence on Labour Providers in Context of NSW and Access to Social Protection This study, designed and directed by one of the authors (Codagnone et al., 2018b), was conducted in support of the European Commission initiative on extending social protection across different forms of employment.12 On the basis of the evidence from this and other studies, the European Commission (2018) has adopted a proposal for a council recommendation on access to social protection for workers and the self-employed. This is a part of the implementation of the European Pillar of Social Rights; we come back to it when discussing current regulatory and policy developments. The sample comprised 8,000 respondents (800 from 10 countries each: France, Germany, Italy, the Netherlands, Poland, Portugal, Romania, Slovakia, Spain and Sweden) and was representative of the online population in each country but stratified as follows: (a) 40% in open-ended full- or part-time work; (b) 40% in NSW (temporary, part- or full-time contract and self-employed); (c) 20% unemployed and looking for a job. Of these 8,000 individuals, 1,893 reported that they generate income from digital platforms, which represent 23% of the total sample but 25% of those in the sample who are not unemployed (7,502). Below, we mostly focus on the employment status of those participating on the platform, the motivations and on perspectives of social protection. On socio-demographic dimensions, this survey broadly confirms the findings of other surveys (younger, highly educated, 50:50 gender balance). In terms of education, 33% of those working with platforms have university degree and above (15% postgraduates). The graph in Fig. 6 shows the breakdown between doing work, selling goods or renting space and doing both activities. The cross-tabulation of participation by employment status is statistically significant and shows a polarisation where two groups are the most represented ones: the self-employed and also individuals with full-time open-ended contract, although not in the proportion as seen in previous survey; this could be explained by the fact that this sample is better stratified and reduces the percentage of fulltime employees (probably over-represented in non-stratified sample). But aside from this difference in numbers, the relatively high share of full-time employees is confirmed. As per the motivation, 55% report that they can work from home, 34% inform that they have more flexibility and only 11% complain that they could not find a regular employment. This latter percentage, however, increases for individuals in lower-income brackets and for certain countries. Italy with 20% and Germany with 4% stand out at the opposite end of the spectrum, but above average, we

12

See all information at http://ec.europa.eu/social/main.jsp?catId=1312&langId=en

98    Platform Economics

Fig. 6.  Renting/Selling, Doing Work, Doing Both.

Fig. 7.  Renting/Selling, Doing Work, Doing Both by Employment Status.

Digital Labour Markets in a Broader Perspective    99

Fig. 8.  Renting/Selling, Doing Work, Doing Both: Self-employed and Full-time Employees. also find both France and Spain, with 15% of respondents reporting that they use platforms to generate income because they could not find regular employment. Moreover, if we consider only those doing work and compare self-employed with full-time employees, the reason ‘could not find a regular job’ is much more frequent among the former. In order to understand relationship between working in digital labour markets and social protection, it is worth presenting some information on the sample as a whole on this issue. The survey asked respondents about their judgement on the adequacy of the social protection they are entitled to and the level of coverage they had. It also explored how worried they were about certain risks (i.e. losing their job, accidents, etc.). Following a previous study (Burgoon & Dekker, 2010), the survey explored two hypotheses: H1. Individuals in temporary employment and possibly in regular part-time employment and self-employment ought to report more perceived insecurity than those in full-time open-ended employment. H2. Individuals in temporary employment and possibly in regular part-time employment and self-employment ought to perceive the level of social protection in their country in less favourable terms as compared with those in fulltime open-ended employment. Both hypotheses are confirmed by the data at descriptive and in a regression analysis conducted on the sample as a whole. Social protection in terms of the six types of benefits analysed in the survey (unemployment, old age, maternity/paternity,

100    Platform Economics sickness, invalidity, accidents/occupational diseases) is considered inadequate by 60% of respondents. Such appraisal differs markedly by employment status, for instance: taking the self-employed and those employed full time on an openended contract the appraisals are specular: 73% of the former consider is not adequate at all (25%) or not very adequate (48%); whereas for the latter this percentage goes down to 42% with as much as 58% considering it very adequate (7%) or fairly adequate (51%). Among those who are in involuntary self-employments the percentage of those considering social protection not very adequate and/or not adequate at all jumps to 86%. For none of the six types of benefits coverage reach 50% of the sample. Moreover, coverage differs widely by forms of employment, for instance: 72% in open-ended full-time employment have access to unemployment benefits but only 22% among the self-employed and 43% among the part-time employed with temporary contract. The unemployed show low level of coverage for the five types of benefits other than unemployment one. For instance, coverage for old-age ranges from only 10% to 23% depending on the forms of employment in which they worked before becoming unemployed. Not surprisingly, a regression analysis focussing only on respondents reporting that they do work for others using platforms show that being self-employed is negatively associated to the judgement of adequacy of social protection.

Evidence on Broadly Defined Market Efficiency Hypotheses In this section, we consider the findings of experimental and quasi-experimental studies that can shed light on some of the economic hypotheses discussed earlier. Is the OLMs world really flat? Horton’s (2010) initial optimistic ‘flat world’ hypothesis was that OLMs would enable global matching, unlimited ‘virtual labour migration’ and international human capital specialisation. Although at descriptive level international flows of digital work seem fairly widespread and even dominant for Upwork, more sophisticated analyses of data suggest that in OLMs the world is not as flat as Horton (2010) predicted when it comes to geographical, cultural and language differences (Agrawal, Lacetera, & Lyons, 2013b; Beerepoot & Lambregts, 2015; Galperin, Viecens, & Greppi, 2015; Ghani, Kerr, & Stanton, 2014; Hong & Pavlou, 2014; Lehdonvirta, Barnard, Graham, & Hjorth, 2014; Mill, 2011); as put it by Lehdonvirta et al. (2014), even in OLMs there is still a high ‘liability of foreignness’. Using data from Freelancers, Mill (2011) showed that when they have no experience in this OLM, contractors from developing countries are less likely to be hired. Another study on Freelancers’ contractors from poor, non-English speaking countries with traditional (religious vs. secular) cultural values and a large time zone difference from employers’ geographical locations such North America and Europe find that they have more difficulty being employed (Hong & Pavlou, 2014). Agrawal et al. (2013b) present results for oDesk that are counterintuitive with respect to those reported above. Focussing on contractors from low-income countries, they find that, all else being equal, those with no experience have a much lower probability of being employed. On the other hand, when contractors from these countries have prior experience, they are disproportionately at an advantage. A different but related finding is that members of Indian Diaspora

Digital Labour Markets in a Broader Perspective    101 who hire on oDesk are more likely to hire workers in India than other employers (Ghani et al., 2014). Lehdonvirta et al. (2014) test and empirically confirms using oDesk data the hypotheses that international digital labour flows are hindered by (i) practical barriers (language differences and time zones) and (ii) the liability of foreignness (more complex work and work involving formal institutions and/ or communication work) and that (iii) foreign contractors are paid less for the same type of work relative to domestic contractors, with this gap being greater in complex work, work that directly involves formal institutions and communication work (interaction between this hypothesis and the previous one). Beerepoot and Lambregts (2015), with data from oDesk, empirically test the hypothesis of wage convergences due to globalised flows. They find that Western contractors earn more than non-Western ones, although when earnings are normalised, using data reflecting the countries of origin’s economic contest, a wage premium for non-Western contractor emerges and documents a limited level of convergence. Non-Western contractors earn relatively more than Western ones in their domestic markets, but this does not seem to drive down the earnings of Western ones. Furthermore, they find that there is no correlation between earnings and skills/ experience and that reputational mechanisms have a greater effect. This suggests that the Upwork claim of being a global digital meritocracy is overstated at best. It is interesting to note that Beerepoot and Lambregts’ (2015, p. 247) claim that it is not uncommon to find in posts the statements such as the following:

This job is not for people from Bangladesh and Pakistan and your bid would be rejected automatically if you are from any one of the mentioned countries.



Business to business appointment setters needed: with previous calling experience Filipinos are preferred.



The client has requested they want a female caller with a British or Australian or New Zealand accent working on the campaign. MEANING UNLESS YOU ARE FEMALE AND UNLESS YOU ARE A KIWI, AN AUSSIE OR BRITISH, DO NOT APPLY!!!!

The authors comment that such forms of discrimination thrive because of the regulatory vacuum in which such transaction takes place. The limits evidenced here are examples of the matching frictions and hiring biases that are further discussed below.

Matching Frictions, Market Inefficiencies and Biases The chapter by Autor (2001) on ‘wired labour’ is cited by many of the authors discussed below; also relevant are labour economic studies of markets with matching functions and search (Petrongolo & Pissarides, 2001, 2006). Probably, the most striking evidence of the existence of frictions and entry-level hiring inefficiencies comes from a field experiment run by Pallais (2013) using oDesk.

102    Platform Economics Adopting the ‘Experimenter as employer’ framework, the author posted a 10-hour data entry task and randomly hired 952 contractors, providing them with a rating when they completed the task (treated group); 2,815 contractors that applied but not hired were used as the control group. Subsequently, the employment performance in oDesk of both groups was monitored. It was found that, considering only those contractors with no prior experience in oDesk, the income for the treated group was three times higher than that for the control group during the two months following the experiment. The employment performance for the treated group can be attributed to the information the author has posted on those who were ‘fictitiously’ hired (since only contractors with no other prior experience were considered in the experiment). The evidence is striking, considering how small the treatment was (a short simple job and a single score out of 5) compared with the size of the effect produced. This clearly suggests important frictions and hiring inefficiencies and biases, especially for entrylevel contractors. The author concludes that the welfare implications are that it had been inefficient not to employ some of the experimental workers; she further argues that OLMs may exacerbate wage inequalities by further skewing work in favour of the most skilled and precluding entry by inexperienced workers. Another three experiments conducted on oDesk by the same author confirm that, all else being equal, referred workers are more likely to be hired than non-referred workers (Pallais & Sands, 2016). It seems that referrals information is used by employers more than all other observable characteristics on which information is fully available in OLMs. In another observational study of oDesk, it emerges that inexperienced contractors affiliated to an intermediary agency (active in using oDesk to mediate between employers and contractors) have substantially higher job-finding probabilities (almost double) and wages (15% more) at the beginning of their careers than inexperienced contractors not affiliated to an agency of this kind (Stanton & Thomas, 2014). This study, besides confirming entry-level frictions, also underscores the importance of this kind of outsourcing agency (as many as 1,100 such firms are active in oDesk) that intermediate between the workers in the OLMs and potential employers. In practice, they act as re-sellers. In other words, for that part of the transaction intermediated by these agencies, oDesk (and other OLMs where this practice exists) ceases to be a 2SM. This seems to confirm the prediction that these new forms of wired labour would require new intermediaries to reduce frictions and increase workers’ productivity (Autor, 2001). This chapter shows that in the case of oDesk these agencies reduce information frictions in the market by screening workers and communicating the results to employers. A typical agency represents a small number of workers, often from the same region or city, and in many cases, they know each other offline. It is worth noting that the presence of these intermediating agencies is observed not only on a similar platform (Freelancers) but also on MTurk. Here, they figure among the top requesters as they aggregate tasks on behalf of their clients to provide a qualityassurance service, on top of MTurk’ services (Ipeirotis, 2010). An experiment run by oDesk shows that algorithmically recommending workers to employers for the purposes of recruitment can substantially increase hiring (Horton, 2015a). Employers with technical job vacancies that received recruiting recommendations had a 20% higher fill rate than the control. Another experiment

Digital Labour Markets in a Broader Perspective    103 run in Elance-oDesk shows that employers were asked for their price/quality preferences before posting their job openings, then these preferences were exposed to would-be workers, a substantial sorting was done by workers and better matches were achieved compared to the control group (Horton & Johari, 2015). These experimental findings are confirmed by observational studies. Traditional ratings are not efficient to reduce frictions and may actually increase them (Horton & Golden, 2015). Another observational study of Elance-oDesk shows the existence of supply constraint frictional effects: spurned invitations affect subsequent match formation (Horton, 2015b). The author shows that sellers are more likely to reject interested buyers when these sellers have more proposals to choose from and, using an instrumental variable identification strategy, argues that this relationship is likely to be causal. When a buyer is rejected by a seller, the latter’s chances of filling his/her request is reduced. This is possibly so because the seller may be pursuing a ‘superstar’ contractor. Market frictions also affect MLMs, such as TaskRabbit, for the delivery of physical tasks (Cullen & Farronato, 2015). In TaskRabbit, demand is highly variable and there is a wide heterogeneity in the tasks posted by the ‘requesters’ and the skills offered by the sellers. The authors find that during the period considered (before change in the TaskRabbit business model): (a) the natural level of efficiency of this market was very modest (although with some differences across different cities); (b) there were clear frictions partly compensated by elasticity in the supply of labour and (c) matching success varies across cities as a function of geographic density (buyers and sellers living close) and task standardisation (buyers requesting homogeneous tasks). Interestingly, they find that when demand exceeds supply, there is no effect on price but rather the supply expands. In other words, contractors work more but the average price remains between $52 and $59 per job. In view of these findings, the platform has been redesigned to increase the efficiency of matching, moving from the original auction model to a new more controlled and standardised business model. In the original model, a buyer could post a task-specific price and then accept the first offer, or ask for bids and review the prices offered by sellers. This move away from the auction business model to more centralisation is part of a growing trend (Einav et al., 2013). After the change in its business model, TaskRabbit basically accepts standardised tasks that are offered at fixed prices.13 So, whereas originally TaskRabbit was presented as the eBay for physical odd jobs, today its ambition is to become the ‘Uber for everything’ (Newton, 2014). Finally, Cullen and Farronato (2015) use their model to estimate the aggregate value of the market for domestic tasks in the United States (considering only 18 cities) at $920 million14 in total. 13

It must be noted that, though focussing on a different type of market such as Freelancers, an analysis of a dataset containing both open auctions and sealed bid transactions shows that the latter attract more bids but the former offer buyers higher surplus (Hong, Wang, & Pavlou, 2014). 14 They estimate the average value per match at $37 and the average number of tasks per requester at 1.23 per month. In addition, they assume that 20% of US households in the 18 cities post requests and that platforms are able to match 80% of them.

104    Platform Economics The contributions above studied and interpreted frictions and hiring inefficiency from a strictly economics and technical perspective. Studying gender hiring in one OLM (name not revealed) other authors find (Silberzahn, Uhlmann, & Zhu, 2014) and discuss (Uhlmann & Silberzahn, 2014) discrimination that is rooted in cognitive heuristics and biases (i.e. conformity under uncertainty) affecting employers judgement and decision making. The main empirical findings are that, controlling for other relevant parameters: (a) female workers are less likely to be hired for stereotypically male jobs (i.e. programming) and more likely to get stereotypically female jobs (i.e. customer service); (b) in the less likely cases in which women are hired for stereotypically male jobs, they are more often paid by the hour rather than by a fixed price of the output (whereas the reverse is true when women are hired for stereotypically female jobs) as a result of a risk-averse choice based on gender stereotypes (i.e. employers are uncertain about women doing a good programming job and this reduces risk by the hour rather than the final output). In the comments to such findings – presented in a separate piece published in a special issue of Behavioral and Brain Sciences on big data and the study of collective behaviour – the authors first observe that OLMs in principle approach very closely a perfect market, given the amount of information available on workers that is searchable and organised algorithmically, which should enable a typical employer to act as perfect Homo Oeconomicus (Uhlmann & Silberzahn, 2014, p. 103) and hire rationally on the basis of skills, merit and value for money; the fact that this does not occur for women with higher skills and better price/quality ratio compared to men is interpreted as a sign of typical heuristic and biases in the face of information overload. When one receives 100 CVs one hour after posting a job he/she faces the situation of assessing many options along many criteria, which is typical of other situations where it is common to rely on social convention heuristics, such as stereotypes, and make potentially biased decision. Cultural stereotypes and confirmation can thus be seen as playing a role in the observed ‘discrimination’ concerning hiring decision and contract types for women (Uhlmann & Silberzahn, 2014, p. 104). The empirical findings fully confirm that stereotypes are inaccurate and cause distorted decisions, since the data show unequivocally that women applying for stereotypically male jobs possess, on average, more domain-relevant skills than their male counterparts (Silberzahn et al., 2014). These mechanisms reproduce themselves inasmuch employers who choose based on gender stereotypes and are satisfied with the output and will never test counter-stereotypical hiring, thus reinforcing and confirming their own biases.

Super Star or Long Tail Effects? The distributional employment and related income effects depend on whether ‘superstar effects’ (leading to income inequality)15 or ‘long tail effects’ (having 15

Superstar effects should produce increased income inequality because employers choose the best workers based on global rather than on local search. When there is a large difference in wages between the global online market and the local physical one (as between higher- and lower-income countries), the superstar effect will drive up the wages of high-quality global online workers, and as a result those of local workers

Digital Labour Markets in a Broader Perspective    105 an equalising impact) prevail? Citing evidence from other online platforms (but not labour market ones), Agrawal et al. (2013a, pp. 14–17) conclude that both results are possible and the evidence on these effects is ambiguous and inconclusive. Yet, in the reviewed studies focussing on OLMs, examples of concentration of work assignments (if not fully blown ‘superstar’ effects) are found and no case of long tail effect is documented (Horton, 2014; Ipeirotis, 2010; Musthag & Ganesan, 2013). In oDesk, for instance, buyers inefficiently pursue oversubscribed (i.e. superstars) sellers (Horton, 2014, 2015b). MTurk is a heavy-tailed market in terms of both ‘requesters’ and ‘turkers’ (Ipeirotis, 2010). The top 0.1% of requesters account for 30% of the dollar activity and 1% of them post more than 50% of the dollar-weighted tasks, and 10% of ‘turkers’ perform 75% of the completed tasks. Such effects, moreover, can be related to barriers to internationalisation seen above and to matching/frictions and hiring bias. Beerepoot and Lambregts (2015, p. 250), for instance, find that Filipino contractors are preferred by many employers for administrative support services, which makes it very difficult for other groups to compete for this task category.

Why Do Firms Hire from OLMs? This is a crucial research question because it refers to the theory of the firm and the possibility that OLMs may increase the contraction in firms’ boundaries in the same way as outsourcing has done since the 1990s. This may impact future development trends in these digital markets. According to the theory of the firm and the transaction costs theory, if OLMs reduce transaction costs, firms should contract more workers in this fashion and this would produce a distributional transfer of work activity from vertically integrated firms to OLMs. Unfortunately, no scientific study on this aspect is available and only a few descriptive data from surveys commissioned by OLMs can be reported. The first survey of 7,000 employers was conducted by oDesk in 2012 (few findings are reported in Agrawal et al., 2013a, p. 12); this survey found that 76% of employers indicated that they hired ‘remote workers’ because they are less expensive. However, 46% selected the answer ‘can get work done faster’, and 31% selected ‘difficult to find talent locally’. A more recent survey on ElanceoDesk (2014) broadly confirms the above findings, although about 60% indicated work being less expensive as a motivation (a bit less than in the previous survey). What is more interesting is that in both surveys between 15% (2012) and 20% (2014) of employers indicated that in the absence of the digital hiring will be driven down. If information asymmetries are present, they may exacerbate the superstar effects. Vertical differentiation in quality may produce superstar effects, whereas horizontal differentiation (variety) may drive long-tail effects (Bar-Isaac, Caruana, & Cuñat, 2012). The long-tail effect, in fact, may occur for workers offering less common and less locally demanded areas of expertise, wages for less common skills may be low because local demand is limited and digitization with access to many different and distant markets could greatly increase the demand for such skills relatively to supply, and hence greatly increase wages.

106    Platform Economics possibility they would have made a traditional local hiring; this means that the boundaries of only one in five firms are affected. The survey of Nubelo (2014) employers report that cheaper labour is indicated as a key reason by 52% of employers, and 30% of employers state that they would make a traditional local hiring in the absence of Nubelo (i.e. firms’ boundaries seem to be more affected). Finally, it is worth noting in both these OLMs that the overwhelming majority of employers are SMEs, which seems to suggest that OLMs make outsourcing more available to this type of firms. Gurvich et al. (2015) provide further insights into the future of digital outsourcing, albeit not empirically but rather from formalised modelling simulation of on-demand work. They model a situation where firms use on-demand work with self-scheduling, meaning that workers are fully autonomous and decide independently how much labour to supply. Their model concludes that under this configuration, compared to a scenario where it is possible to dictate to workers when they must work, the firm has lower profits, and customers have a higher chance of not being served. Furthermore, the modelling foresees that, when demand is volatile, self-scheduling results in lower service levels during high demand periods. This analysis clearly points to the limits of outsourcing and relates to the above discussion about the firm’s boundary. It can also be seen, on the other hand, as indirectly supporting the claim that digital labour markets must exert strong control over on-demand workers and possibly influence their patterns of work if they are to be profitable and used by firms.

Net Aggregate Effects The evidence available does not warrant any conclusion with respect to the net welfare effects that online or MLMs and services have. The only source dealing with these aspects for OLMs is the well-reasoned and balanced theoretical discussion presented in Agrawal et al. (2013a, pp. 23–25). There are two kinds of possible aggregate welfare effects: (1) Increased efficiency in overall labour market matching (i.e. increased pool of workers and employers, lower transaction and search costs). (2) Increased production efficiency due to lower coordination costs. For the first effect, the evidence reviewed suggests that online and MLMs still have considerable frictions and inefficiencies, and some intangible obstacles hamper the full impact of breaking geographical barriers. Agrawal et al. (2013a) recognise that these markets must further improve their design and two-sided strategies to improve matching because: (1) OLMs reduce search costs but increase heterogeneity in the pool of workers and employers, and, consequently, in the skills and tasks to be matched. (2) This may compound the potentially negative effects of the lack of soft- and/ or high-bandwidth information in OLMs about both job seekers and prospective employers (Autor, 2001).

Digital Labour Markets in a Broader Perspective    107 (3) Lower coordination and lower outsourcing costs through these new labour markets may increase production efficiency, yet this is still a speculation since no empirical evidence is found on this aspect. So far, we have focussed on OLMs for the simple reason that evidence on MLMs is very limited and is briefly discussed below for distributional effects only.

Distributional Effects in MLMs Compared to OLMs, the MLMs by definition cannot have global geographical effects, for they are localised. Providing personal and home services obviously does not have any effect on firms’ boundaries. Markets such as FieldAgent or Wegolook, which provide services to business, might in principle have these effects, but no empirical evidence was found on these types of services. Musthag and Ganesan (2013) show that one platform for mobile services is heavy tailed, that is, less than 10% of workers account for more than 80% of the activity generated by the platform. On the other hand, a survey of TaskRabbit workers documents a different kind of distributional effect (Thebault-Spieker et al., 2015): low socio-economic status (SES) areas are relatively less serviced by TaskRabbit. They conclude that more research is needed to document whether their results can be generalised and can support the conclusion that lower SES groups and/or neighbourhoods have less chance of benefiting from the sharing economy.

Digital Labour Markets in a Broader Perspective Evidence is still needed to fully size the dimension of digital labour markets, although we can conclude that it is a statistically non-marginal and detectable phenomenon. On the other hand, its growth in the last five years has been phenomenal, and if growth continues at this fast pace, these new markets could encroach on traditional and long-term forms of employment (Einav et al., 2015, p. 20). The questions we aim to address in this section in a more speculative fashion are: What are the possible futures of digital labour markets putting them in relation with the future of work themes? and How do they fit in the broader trend of the growing relevance of NSW?

Technological Trends: Digital Labour Markets and the Future of Work As recounted by Schwartz (2015) in a piece tellingly titled ‘Human pretending to be computers pretending to be human’, Wolfgang von Kempelen (1770) in Vienna presented Empress Maria Theresa a sort of robot that could beat humans at playing chess; he called it the ‘Turk’. The ‘Turk’ toured Europe and evoked contrasting responses, as some were ready to admit and welcome that the machines were surpassing humans, but many opposed this view. Although the Turk included a ‘labyrinth of levers, cogs and clockwork machinery’, obviously it was not using any algorithm and was operated by a person hidden inside. The first digital platform for the trading of micro-tasks, AMT, was presented as

108    Platform Economics a new technology of ‘humans-as-a-service’ as opposed to ‘software-as-a-service’ and as a new form of ‘artificial intelligence’ (Irani, 2015, p. 225). More recently, engineers have further brought this idea by designing Soylent, ‘a word processor with a crowd inside’ that uses workers from MTurk to create a ‘human-powered’ editor and writing assistant ‘inside’ MS Word (Bernstein et al., 2015). So, one could wonder as to what extent the future will be about robots supplanting humans, or conversely about digital transformations turning humans into ‘robots’. Past and current technological trends are projected into various descriptions of the future of work in a crucial policy-relevant debate. The future of work is seen by many economist as the result of unstoppable technological change, as in the narrative of the ‘flat world’ (Freeman, 2008; Friedman, 2005) or in the ‘Skill Biased Technological Change’ (SBTC) hypothesis (Autor, 2013; Autor & Dorn, 2013; Autor et al., 2003). Computerisation and robotisation as part of the so-called ‘second machine age’ (Brynjolfsson & McAfee, 2012, 2014) is a pivotal hypothesis and narrative in this debate. Using the SBTC tasks taxonomy, it has been estimated that in the United States and Europe the risk of job computerisation stands, respectively, at 47% (Frey & Osborne, 2017)16 and between 40% and 60% (Bowles, 2014); a more recent OECD working paper, however, estimated that in OECD countries on average 9% of jobs are at high risk of being automated (Arntz, Gregory, & Zierahn, 2016). Given this debate, it is natural to ask why digital labour markets trading routine tasks and micro-tasks are booming? Not only MTurk but also OLMs, such as Upwork, are trading tasks into which jobs have been broken down. The most traded tasks in Upwork are fairly routine ones, such as for administrative support as a category, or accountants as a position (growth rate, respectively, 37% and 43%). The next question is whether these markets are here to stay, or is this a transitory phenomenon? Using insights from the limits of ‘wired labour’ (Autor, 2001, 2008) and from the theory of the firm, one may expect that outsourcing by firms may be hampered by the matching inefficiency of digital labour market, or alternatively by intervening regulation. Advances in computerisation and eventually in robotisation and their gradual cost reduction (as occurred with traditional Information Communication Technology [ICT]) may lead to more tasks being carried out by machines. In this scenario, firms may no longer need to outsource tasks to digital labour markets that could be internalised by using robots. On the other hand, robotisation may advance slower than expected (in terms of effectiveness) or may be hindered by intervening regulation. In this case, more space would be left for digital labour platforms. Finally, even under high robotisation, firms may still outsource (for various reasons) so that digital labour platforms may retain a complementary role. To sum up, how large digitally intermediated on-demand work is now, or may become in the future, remains the subject of debate and controversial forecasts (Zumbrun & Sussman, 2015).

16

Note, however, that this paper has been heavily criticised in the labour economics literature.

Digital Labour Markets in a Broader Perspective    109 Yet, going back to the first question, there are alternative answers such as excessive costs of computerisation, especially for SMEs, fluctuating demand for tasks, limits to the possibility of routinising work, especially labour cost saving made possible by institutional change such as work de-standardisation and regulatory arbitrage. This latter hypothesis links the emergence of digital labour markets to institutional changes and not only to technology, an issue we discuss next.

Not Simply Technology Between the mid-1990s and 2010 a clear job polarisation trend (U-shaped work force with the relative decline of middle-skilled jobs and increase of high- and low-skilled ones) has emerged in most OECD countries: between 1995 and 2010, routine jobs (i.e. accountants) fell 12 points (from 53% to 41%), while high-skill abstract jobs (i.e. designers) increased 10 points from 28% to 38% and nonroutine manual jobs (i.e. drivers) increased three points from 18% to 21% (OECD, 2015a, p. 29); this trend is seen as mostly produced by technological change and is expected to increase in the coming years. As argued on the basis of an econometric analysis by OECD (2015a, pp. 147–152), technological trends and RBTC hypothesis tell only one part of the story. Job polarisation is clearly associated with de-standardisation of labour contracts and there is no conclusive evidence demonstrating what caused what between technology and institutional change. What is clear from the data is that routine jobs based on standard contracts (i.e. accountants) have been substituted by routine jobs in NSW forms and this is exactly also happening in OLMs. If the disappearance of accounting jobs in the middle of the jobs hierarchy was entirely driven by technology, then such jobs could not have been substituted by individuals working under NSW conditions having the same skills as those laid off; this is what actually occurred as shown by the data. Digital labour markets are ‘moving frontier at which machines still fail to carry out tasks by employing humans to help machines to learn’ (OECD, 2016, p. 22). This is the case of AMT with its HIT. It has been observed, how in the past three decades or so each new wave of digital innovation gave rise to contrasting narratives about the future of work (Baldry, 2011; Holtgrewe, 2014; Howcroft & Taylor, 2014). Such unilateral focus on technology obfuscates the importance of institutional changes in labour markets such as, for instance, the de-standardisation of work and the emergence of NSW, which is statistically associated with inequality and poverty (OECD, 2015b, pp. 152–170) and is, thus, related to the ‘new inequality’ debate (Atkinson, 2015; Bernhardt, 2014; Kuttner, 2013, 2016; Standing, 2011; Summers & Balls, 2015).

Non-standard Work and its Effects In order to consider ambivalence and challenges of flexibilisation trends, it is worth contrasting the potential benefits and risks of these forms of works. Flexible non-standard working practices have various positive aspects for labour market actors. Non-standard employment has been identified as a means to improve work opportunities, notably for women and migrant

110    Platform Economics workers, and for countering high levels of unemployment by creating new, flexible jobs (ILO, 1997). Firstly, flexible contracts give employees the opportunity to reveal or signal their productivity to their prospective employers. This ‘stepping stone’ interpretation of non-standard employment (de GraafZijl, van den Berg, & Heyma, 2011) suggests that such contracts may help to reduce informational asymmetries and improve matching between jobs and candidates (Ichino, Mealli, & Nannicini, 2008). Secondly, non-standard contracts may be preferred because of the opportunity afforded to the individuals for greater flexibility. Given that employment security rests increasingly on individuals taking responsibility for their current and the future human capital development (Urtasun & Núñez, 2012), gaining firm-specific human capital via the rigidities of a permanent full-time contract may be less attractive than was the case formerly. Thirdly, flexible forms of employment may suit those individuals who desire to balance their working and non-working lives. On the other hand, job insecurity and precariousness can negatively impact on physical and mental health (Burchell, 2009; Herbig, Dragano, & Angerer, 2013; Paul & Moser, 2009; Virtanen et al., 2005). Precarious jobs may become ‘traps’ as opposed to ‘bridges’ into secure work and reduce social mobility (Cahuc & Kramarz, 2005; Gash, 2008; ILO, 1997). A large study of Europeans aged 20–35 years shows, for instance, that temporary work is a choice among the younger group (aged 20–25 years) but tends to become a trap for the those aged 31–35 years who cannot find better employment (Nunez & Livanos, 2014). Trends towards work becoming more flexible have also been associated with growing inequality. A study using EU-SILC data for 24 European countries shows that temporarily employed have a higher poverty risk than permanent workers, mainly due to lower wages (Van Lancker, 2012). The fiscal costs would in the end would increase for governments dealing with the social costs of job insecurity (Adams & Deakin, 2014). A large proportion of individuals in NSW are not fully covered or not covered at all by social protection systems (ILO, 2016, 2017; Matsaganis, Özdemir, Ward, & Zavakou, 2016; Spasova, Bouget, D., Ghailani, D., & Vanhercke, 2017), especially because there is a substantial gap between statutory and effective accession to benefits (building up entitlements): ‘Even if non-standard workers are covered by a social protection scheme and thus formally have access to the related benefit, they may not have effective access to it because eligibility criteria are set in such a way that they have major difficulties meeting these’ (Spasova et al., 2017, p. 14). According to a calculation presented in Matsaganis et al. (2016), in Europe 13% of all those employed aged 15–64 years are at risk of not being entitled to unemployed benefits, for they are in one form of NSW and for sickness and maternity benefits the share is 8%. Most of the potentially negative effects briefly reviewed above are certainly relevant for providers of labour services in digital labour markets, given that most of these operators are not fully regulated, and so providers enjoy little rights and no social protection. This observation leads us to the final section where legal disputes and policy and regulatory issues are discussed.

Digital Labour Markets in a Broader Perspective    111

From Legal Battles towards Regulation and Policy? Even more than for other ‘sharing platforms’, the emergence of digital labour markets has ignited countless court litigations, first in the United States and more recently in Europe, stimulating a growing body of interdisciplinary literature dealing with the legal, regulatory and policy dimensions of the issue (Aloisi, 2015; Cherry, 2011, 2016; Cherry & Aloisi, 2016; De Stefano, 2016; Dokko, Mumford, & Schanzenbach, 2015; Dubal, 2017; Felstiner, 2011; Graham, Hjorth, & Lehdonvirta, 2017; Hagiu & Biederman, 2015; Harris & Krueger, 2015; Hill, 2015a, 2015b; Minter, 2017; Posen, 2016; Prassl & Risak, 2016; Ravenelle, 2017; Sprague, 2015; Stewart & Stanford, 2017; Todolí-Signes, 2017). As regulators and policy makers have been slow to react, courts and judges are increasingly called to rule, especially on the status of providers in digital labour markets: Are they misclassified as contractors and are rather employees? Given the prominence and importance of these litigations, in this section, we first consider US litigation cases and then European ones. Next, we review the ongoing debate to conclude by considering the EU-level regulatory and policy framework.

US Legal Disputes In the United States, a number of litigations have been brought to the courts concerning the possible misclassification of workers as contractors as well as other violations of the Fair Labor Standard Act (FLSA). Table 3 provides a selective list of such cases: While the litigations concerning Uber and Lyft have attracted the most of the media attention, the table shows that several other OLMs were concerned, including Crowdflower, for violation of FLSA related to minimum wages. In key litigations concerning Lyft and Uber (Cotter v. Lyft Inc., 60 F. Supp. 3d 1067, N.D. Cal. 2015, and O’Connor v. Uber Technologies Inc., No. C-13-3826 EMC, 2015 WL 1069092, N.D. Cal. Mar. 11, 2015), judges pointed out that drivers possess no special skills, their input is essential to the businesses, they are monitored and terminated if they do not comply with rules set by the two companies. In the Uber case the company claimed to be just a software company that was deemed by the court; however, this line of defence was ‘fatally flawed in numerous respects’ as it focused exclusively on the mechanics of the platform rather than on the substance of what Uber actually does (O’Connor, 2015 WL 1069092, at *6). The plaintiff, in fact, cited the Drivers’ Handbook where it is clearly written that drivers should accept all ride requests and that if a driver rejects too many trips, he/she will be investigated and possibly terminated. Uber argued that it never performs inspections. Many cases have been settled to avoid court rulings that change regulatory framework for the entire world of OLMs and MLMs. On the other hand, despite these settlements there is ‘no guarantee that the Internal Revenue Service, the NLRB or another governmental regulator will not determine that these workers are employees’ (Cherry, 2016, p. 7); in other words, the issue of classification remains open in the United States. There are complex multifactor tests defined by the law and applied to decide whether a person is a contractor or an employee (Cherry, 2011; Harris & Krueger, 2015; Sprague, 2015); the Uber

112    Platform Economics Table 3:  Selected Litigation Cases in the United States. Platform

Object of Litigation

Status

Crowdflower

FLSA violations with respect to minimum wages

Settlement reached (payment of $583,000), parties agreed an amendment denied by the court

Handy

Misclassification class-action lawsuit

Unsuccessful mediation attempted will go towards arbitration

Homejoy

Misclassification class-action lawsuit

Class claims dismissed (Homejoy ceased operations)

Lyft

Employee benefits, cost Settlement for $12 million on reimbursements 27 January 2016; as part of settlement, termination of drivers will be subject to due process rights

Postmates

Class-action for violation of FLSA

Ongoing

Uber 1 (vs O’Connor)

Employee benefits, cost reimbursements; overtime under FLSA

Settlement for $12 million on 21 April 2016; as part of settlement, termination of drivers will be subject to due process rights

Uber 2 (vs Ehret) Employee benefits, cost Ongoing reimbursements Uber 3 (vs Mohamed)

Employee benefits, cost Ongoing reimbursements

Sources: Compiled from Cherry (2016), DeAmicis (2015), Kessler (2015), Madden (2015) and Smith and Leberstein, (2015).

and Lyft litigations, though not producing regulatory effects, have ascertained the control exerted by the Uber and Lyft drivers, highlighted the associated liability and deemed their claim that they are just software companies simply untenable. As a result of these litigations, in the United States some OLMs and MLMs have started reconsidering their position with regarding labour issues, redrafting their terms of agreement and reclassifying the individuals performing work (Bensinger, 2015a, 2015b; Cherry, 2016; Lang, 2015; Smith & Leberstein, 2015; Stokes et al., 2014). A number of markets have already decided to adopt fairer practices and self-regulation codes. A few examples are provided below.

Box 8: Changing Practices by Digital Labour Markets in the United States. •  T  askRabbit, for instance, since 2015 has set up a wage floor: It is not possible to earn less than $12.80 an hour, which is higher than any minimum wage in the United States (Quittner, 2015).

Digital Labour Markets in a Broader Perspective    113

•  M  unchery (food preparation and delivery) hires their workers as employees (O’Donovan, 2015), as do the house-cleaning services platforms Qii (Koso, 2015) and MyClean (Roose, 2014a), the valet parking service Luxe (Lien, 2015), the mailing company Shyp (Alba, 2015) and the food delivery start-up Sprig (Bensinger, 2015a). •  Instacart has announced that it has reclassified some of its independent contractors as employees (Bilis, 2015). •  Some of these platforms are doing very well economically, despite the alleged additional health insurance and social contribution costs of switching from contractors to employees (Roose, 2014b). Even Uber (2018), that has been facing many litigation cases also in Europe, as part of its public relation strategy has taken a new position launching new driverfriendly App17 and releasing a White Paper on social protection. The interesting thing in this chapter is that Uber proposes ‘to ‘design our social institutions to be more neutral to working style’ (italics emphasis added). It is worth quoting here: Instead of trying to shift everyone into the traditional model, which is exclusionary by its very nature, we believe a better path is to design our social institutions to be more neutral to working style. A system rewarding all forms of work would not only provide greater protections for all, it would offer more opportunities and a wider access to work, and have positive effects on the economy. According to this White Paper, Uber (2018, p. 24) has a positive vision for work in Europe. We see a future with: •

• • •

17

Flexibility – Everyone can work in a way that suits them, and vary their portfolio of work to suit their needs at their discretion, including at different stages of life. Everyone can manage their life with ease and move freely between different modes of work. Opportunity and access – Everyone can reliably find and keep quality, safe work, free from discrimination, and be able to maintain a good standard of living. Security – Everyone has access to a broad set of reliable and affordable social protections. Growth – Everyone is socially mobile and can access lifelong learning and development opportunities.

See https://www.uber.com/newsroom/new-driver-app/.

114    Platform Economics This proposal obviously attempts to redesign social institutions according to Uber’s vision of what social protection should look like, in line with their own corporate social responsibility commitments for the future. Recent attempts at optimistic scholarly accounts of the contribution of Uber include how: Uber has been helping residents of Cairo and Alexandria navigate their challenging cities while offering a new form of work that provides an alternative and/or supplementary source of livelihood for many (Rizk, 2017, p. 21); and adaptability has high value to individuals who have selected into the Uber platform. Our expectation is that technology will enable the growth of more Uber-style work arrangements. While such arrangements may have important downsides relative to the traditional careers they supplant or supplement, we expect that flexibility will be an important source of value in such arrangements (Chen et al. 2017, p. 45) On the other side of the pendulum, Cook et al. (2018, Abstract) argue that ‘the gig economy grows and brings more flexibility in employment, women’s relatively high opportunity cost of non-paid-work time and gender-based preference differences can perpetuate a gender earnings gap even in the absence of discrimination’, while Landier et al. (2016, p. 1) argue that ‘the key feature of the French Uber driver partner population is their young age, which makes them particularly exposed to unemployment risk’. Three more studies address more specific aspects of ride-sharing markets. Hall et al. (2017, p. 1) argue that ‘driver supply of labour to ride-sharing markets is highly elastic, most likely because drivers face no quantity restrictions on how many hours to supply and new drivers face minimal barriers to entry’. Castillo et al. (2018, Abstract) argue that raising prices, either by keeping them consistently high or ‘surge’ pricing only at high demand times, brings demand back under control and avoids these catastrophic failures. Banning (2016) surge pricing would thus likely result in always-high prices. Alternative solutions would undermine ride-hailing’s brand promise. Lastly, Angrist et al. (2017) show that ride-hailing drivers gain considerably from the opportunity to drive without leasing. As a contrast to these narratives, a piece in the Harvard Business Review singled out Uber as an example of how behavioural economics and nudges should not be used (Gino, 2017). The author argues and shows, in fact, that the company uses behavioural nudges in combination with algorithmic management to push drivers to pick up more fares.

Digital Labour Markets in a Broader Perspective    115 Legal Disputes in Europe In October 2016, a UK employment tribunal dismissed the notion that Uber merely coordinates self-employed workers.18 According to the judgement, it would be impossible for workers to grow their businesses through Uber unless ‘growing their business simply means spending more hours at the wheel’. The ruling said, ‘The notion that Uber in London is a mosaic of 30,000 small businesses linked by a common “platform” is to our mind faintly ridiculous’. It also stated that by using the rating system, the platform subjected workers to a performance management/disciplinary procedure, going beyond what is allowed in coordinating mere self-employed workers. The platform was forced to recognise two drivers as employees and guarantee them a minimum wage. In France, the loi Thévenoud of 1 October 2014 amends the regulation of passenger transport services in the Transport Code (code des transports) and banne services such as UberPop. Uber challenged the validity of the law as being contrary to the freedom to conduct a business, but this was rejected by the Conseil Constitutionnel.19 In June 2016, the Paris criminal court ordered Uber to pay €800,000, with half of the fine suspended, and found Pierre-Dimitri Gore-Coty, director for Europe, Middle East and Africa, and Thibaud Simphal, the company’s manager in France, guilty of deceptive commercial practices and being accomplices in operating an illegal transport service. On 20 December 2017, the Grand Chamber of the Court of Justice of the European Union (CJEU) held that Uber’s UberPOP ride-sharing service is a ‘service in the field of transport’ within the meaning of Article 58(1) of the Treaty on the Functioning of the European Union (ECJ, 20 December 2017, Case C-434/15, Asociación Profesional élite Taxi v. Uber Systems Spain SL). The case was: Barcelona Taxi versus Uber (Case C-434/15, Asociación Profesional Elite Taxi). While the ruling is not on drivers’ status, some of the motivations in the opinion of the Advocate General (AG) have implications on that matter. The AG starts by recognising the politicisation of the issue: Although the development of new technologies is, in general, a source of controversy, Uber is a case apart. Its method of operating generates criticisms and questions, but also hopes and new expectations. In the legal field alone, the way Uber works has thrown up questions concerning competition law, consumer protection and employment law, among others. From an economic and social standpoint, the term ‘uberisation’ has even emerged.

18

Courts and Tribunals Judiciary _ Y Aslam, J Farrar and Others v. Uber. Retrieved from https://www.judiciary.gov.uk/judgments/mr-y-aslam-mr-j-farrar-and-others-v-uber/. Accessed on June 15, 2017. 19 Décision no. 2015-484 QPC du 22 septembre 2015. Retrieved from http://www. conseil-constitutionnel.fr/conseilconstitutionnel/francais/les-decisions/acces-par-date/ decisions-depuis-1959/2015/2015-484-qpc/decision-n-2015-484-qpc-du-22-septembre2015.144387.html

116    Platform Economics This request for a preliminary ruling therefore presents the Court with a highly politicised issue that has received a great deal of media attention.20 The AG’s opinion, while not legally binding, contains many interesting considerations concerning the legal status of Uber, and online platforms more generally, including the implications for employment law. The AG with a crisp logic shows that Uber cannot be considered a sharing or collaborative platform because the drivers offer passengers a transport service to a destination selected by the passenger and, accordingly, are paid an amount, which far exceeds the mere reimbursement of expenses incurred. The claim of simply matching supply and demand is dismissed, for it creates its supply, establishes strict rules concerning the essential characteristics of the supply, organises and control how it works. What the AG states on decentralised technology-based control is telling: While this control is not exercised in the context of a traditional employer–employee relationship, one should not be fooled by appearances. Indirect control such as that exercised by Uber, based on financial incentives and decentralised passenger-led ratings, with a scale effect, makes it possible to manage in a way that is just as – if not more – effective than management based on formal orders given by an employer to his employees and direct control over the carrying out of such orders.21 While the final ruling simply stated that Uber is not a technology company but rather a transport company and included no prescription on the employment status of drivers, the way it described and ascertained control implies that under most European national legislations and related testes of employment, drivers would be de facto employees. In Italy, in the litigation brought by Foodora’s riders, the Court of Turin ruled in a different way compared to the UK and France litigations and also to the CJEU. The court rejected an appeal by six riders of Foodora who, after having participated to a mobilisation for their rights, were terminated by the platform.22 The judges accepted the platform’s claim that the riders were self-employed and could be terminated any moment, and thus could not benefit of the right to selforganise as would be done by employees. On the other hand, this sentence may become irrelevant in June 2018, the Vice-Premier of the new Italian government Luigi Di Maio has announced a new law that would force Foodora and other operators to hire labour providers as employees. 20

Opinion of Advocate General Szpunar of 11 May 2017, Case C-434/15, Asociación Profesional Elite Taxi v. Uber Systems Spain SL, ECLI:EU:C:2017:364, para. 1. 21 Para 52. 22 See, for instance, P. Cruciatti ‘Cosa ci dice la sentenza di ieri su Foodora’, Il Post, 12 April 2018. Retrieved from https://www.ilpost.it/2018/04/12/foodora-sentenza/. Accessed on May 2018.

Digital Labour Markets in a Broader Perspective    117 The Debate As the litigation cases were ongoing, both law scholars and economists have taken sides. The typical free-market libertarian thinking is that contractors should be left as such to avoid curbing innovation and labour market efficiency (Koopman et al., 2015; Sundararajan, 2014). One common position is that one could not force existing regulation on a new phenomenon such as Uber, but rather to use a metric-based ‘experimental regulation’ (Posen, 2016). This novel idea of using metrics and fine tune experimentally the regulation to the practices has been advanced in many other commentaries and for other operators, for instance, also for Airbnb, where it was called algorithmic regulation (Quattrone et al., 2016). It is in sharp contrast, however, with the fact that operators are not willing to provide access to the data registered automatically by the underlying technological platforms and applications. At the opposite end, the claim of being technology companies and the issue of strict control on providers is underscored as the basis of regulating labour relations (Ravenelle, 2017). The rise of digital labour markets calls for new initiatives in social policy because it shifts more of the burden of economic risk onto workers even while removing gig workers from many of the employment-bound newdeal-era social insurance programmes (Gerald, 2014). Others, after reviewing the litigations, conclude that a new special labour regulation is needed (TodolíSignes, 2017). According to Stewart and Stanford (2017), the options available to policy makers and regulators include the following: enforcement of existing laws; clarifying or expanding definitions of employment; creating a new category of independent worker; creating rights for workers not employees (i.e. regardless of contractual and employment status) and reconsidering the concept of an employer. It is claimed, however, that doctrinal analysis and employee category must be more attendant to workers’ realities, for not all would want to become employees and some are genuine self-employed (Dubal, 2017). Another aspect that has been raised is the possibility of labour providers to associate and defend their rights, which, for instance, the Italian Court ruling on Foodora’s riders denied. Freedom of association and the effective recognition of the right to collective bargaining are among the workers’ fundamental rights recognised by International Labour Organisation, whereas among the crowdworkers of AMTs some form of workers’ collective organisation has emerged (Irani & Silberman, 2013), these initiatives suffer from some of the typical problems of online activism (Salehi et al., 2015). Furthermore, crowdworkers are in strong competition with others for task allocation, which may reduce cooperation and increase incentives for opportunistic behaviour. Finally, being subjected to reputational ratings, they may avoid engaging in collective behaviour, for this may damage their reputation (De Stefano, 2016, p. 9). Hence, de facto, they are deprived of any forms of representation and bargaining. In Chapter 1, building on the work by Hagiu and Wright (2013, 2015a, 2015b, 2015c), we anticipated the issue whether control over the exchanges between the two parts of a ‘platform’ would imply that some of the commercial operators of the so-called sharing economy are not 2SMs but rather some form of vertically

118    Platform Economics integrated firms. Court rulings and evidence from some empirical studies unequivocally confirm that some digital labour markets strictly control both the trade and the way the providers must operate. Control is crucial when dealing with labour platforms. There are various reasons for this: typical matching frictions, the heterogeneity of tasks/contractors/employers, prominence of on-demand and time-sensitivity (i.e. Uber) and problems of coordination of multiple contractors. Obviously, control is maximised in vertically integrated firms to ensure consistency, speed, timely delivery, coordination and scale. However, control has a cost: It can make independent contractors into employees, which increases costs between 25% and 35% (Hagiu, 2015; Hagiu & Biederman, 2015, and possibly more in the European context). Lower costs should be associated with less control, although some of these platforms seem to be striving to minimise costs and maximise control, almost to the level typical of a vertically integrated firm. So, a new category of dependent contractors, as intermediate between employees and pure contractors, has been proposed as a middle-ground solution. For this new category, digital labour markets would pay some of the social protection costs but not as much as if they were reclassified as employees. This proposal attempts to find a middle ground to address social concerns without imposing a jump on the costs that may make digital labour markets no longer competitive and economically viable. Creating an intermediate category of workers such as dependent contractors or dependent self-employed persons implies to identify suitable definitions. Protection for workers in intermediate categories and the tests for applying them also change significantly among national regulations. A new topic in the regulatory debate emerged in 2015 around the proposal that goes under the name of portability (Berg, 2016; Harris & Krueger, 2015; Hill, 2015a, 2015b; Strom & Schmitt, 2016). As effectively summarised by Berg (2016, p. 2), in basic terms the proposal consists in creating ‘individual security accounts to protect the worker as they move from gig to gig’. Benefits (wage insurance, health insurance, and disability and injuries insurance) should be designed universally and not being tied to specific employers (Strom & Schmitt, 2016, p. 14). The final employers would have obligations similar to those with regular workers or they may share contributions with digital labour markets that could pay half of them (Harris & Krueger 2015). This is also the idea inspiring (more in general for NSW and also for those working in the platform economy) the earlier cited European proposal for a council recommendation on access to social protection for workers and the self-employed (European Commission, 2018), to which we come back at the very end of this chapter. The enactment of some form of regulation to establish the proposed portability of benefits would already represent a positive step forward to ensure more dignified conditions for workers in digital labour markets; yet, they are blatantly not sufficient. We have shown that earnings are at times too low in the absence of any minimum wage rules, the flow of work is unstable and no employment benefits exist, there are clear information and power asymmetries, no protection against privacy violations, and various forms of information- or reputation-based ethnic and gender discriminatory mechanisms occur unregulated. A few authors have proposed wider and more comprehensive and regulatory approaches

Digital Labour Markets in a Broader Perspective    119 (Berg, 2016; Sprague, 2015; Strom & Schmitt, 2016). These include minimum wages, the need to recognise overtime and business expenses, support to workers’ self-organisation and unionisation, changes in the organisation of work, and various other provisions. EU-level Context and Developments.  There have been various regulatory and policy developments in many of the EU28 countries that would be beyond the scope of this final paragraph to review.23 We conclude instead by drawing the contours on EU-level general framework for NSW and considering initiatives specific to work in digital labour markets. We first reconstruct briefly the introduction of labour flexibility by EU directives to conclude with reviewing initiatives addressing the so-called ‘collaborative economy’ adopted by both the European Commission and the European Parliament. EU-level Flexibilisation of Work.  NSW arrangements have been introduced and regulated widely at EU level with the three key directives for part-time (1997), fixed-term (1998) and temporary agency contracts (2008).24 As noted (Peers, 2013), the spirit of these directives was to protect both atypical workers directly from abusive conditions of employment and workers with standard employment contracts indirectly from being undercut by atypical workers. They represent a dual approach aimed to liberalise NSW rules and remove barriers while at the same time providing some protection to workers employed in these forms of employment (they contain provisions banning, in principle, discrimination against atypical workers as compared to standard workers). An appraisal of these measures yields mixed results (Deakin, 2014). Many exceptions have been introduced, and the way some member States have implemented the directives risks nullifying the protective provisions; transition is facilitated from standard work to NSW, but not vice versa, so that these directives seem to have perpetuated labour market dualism. A comparative study of labour market developments in the United States and the EU (DiPrete, Goux, Maurin, & Quesnel-Vallee, 2006), for instance, already done by the middle of the previous decade, highlighted that the precarisation of work in Europe was a more widespread trend than in the United States with growing numbers of insecure jobs where low-skilled workers were concentrated. Therefore, the proposal advanced by Hagiu is probably reasonable for the United States where the labour market situation is relatively 23

This mostly draws from the most updated overview of policy and regulatory developments concerning the issue of workers of the sharing economy in Europe that is provided in a report published by the European Agency for Safety and Health at Work (Garben, 2017.). 24 The EU has adopted three measures concerning ‘atypical’ workers: (1) a Social Partners’ Agreement on part-time work (Directive 97/81 concerning the framework agreement on part-time work concluded by ETUC, UNICE and CEEP [1998] OJ L14/9; extended to the UK by Directive 98/23 [1998] OJ L131/10; (2) a Social Partners’ Agreement on fixedterm work (Directive 1999/70 concerning the framework agreement on fixed-term work concluded by ETUC, UNICE and CEEP [1999] OJ L175/43); (3) a directive on temporary agency work known as the ‘agency work Directive’ (Directive 2008/104 [2008] OJ L327/9. Member States had to apply this Directive by 5 December 2011: Art 11(1)).

120    Platform Economics more binary (0 = contractors; 1 = regular employee) compared to Europe. In this respect, any new regulatory proposal concerning workers in digital labour markets should take into account this aspect and consider the differences existing compared with the US situation. The EU Agenda on Collaborative Economy.  In June 2016, the European Commission presented a European agenda for collaborative economy, a non-binding act of soft law that aims at adapting and interpreting the existing regulation, reassuring rights and obligations of various subjects (European Commission, 2016b). As many other Commission pronouncements, this communication tries to make everyone happy. On the one hand, it states that collaborative economy creates new opportunities for consumers and entrepreneurs, on the other hand, it warns about regulatory problems related to the blurring of established lines between consumer and provider, employee and self-employed, or the professional and non-professional provisions of service. In terms of the labour issue, the communication notes the possibility that when a collaborative platform exerts a high level of control and influence over the provision of the underlying service, the underlying relation may be one of employment. Parameters of control are setting the price, terms and conditions on the user–provider relation, platform owns key assets. Obviously, the last condition undermines the case. Many platforms exert full control without owning assets. Not surprisingly, the communication leaves to a case-by-case approach for determining whether a de facto employment relation exists. European Parliament Resolutions.  The European Parliament has been pushing to bring providers in digital labour markets more firmly into national social security systems amid concern that a changing labour market might leave them out from entitlements and benefits. A cross-party MEP coalition seems to be seeking new regulations to protect crowdworkers. In January 2017, the Parliament voted back a report calling for better worker protections in crowd employment.25 The resolution was eventually approved in May 2017 by the Internal Market Committee,26 and is to be voted by the full house at the 12–15 June 2017 session. The resolution of 19 January 2016 on a digital single market act calls on the member States to ensure that employment and social policies are fit for the purpose of digital innovation, entrepreneurship and the growth of the sharing economy and its potential for more flexible forms of employment, by identifying new forms of employment and assessing the need for the modernisation of social and employment legislation so that existing employment rights and social welfare schemes can also be maintained in the digital world of work. While it recognises that the provision of social security is a member State competence, it asks the 25

MEPs demand social protections for Uber, Deliveroo workers – POLITICO. Retrieved from http://www.politico.eu/article/meps-demand-social-protections-for-uberdeliveroo-workers/. Accessed on June 15, 2017. 26 Internal Market Committee calls for EU strategy on the collaborative economy _ News _ European Parliament. Retrieved from http://www.europarl.europa.eu/news/en/press-room/ 20170503IPR73223/internal-market-committee-calls-for-eu-strategy-on-thecollaborativeeconomy. Accessed on June 15, 2017.

Digital Labour Markets in a Broader Perspective    121 Commission to identify and facilitate exchanges of the best practices in the EU in these areas and at international level. On 15 June 2017, the European Parliament by a strong majority adopted a resolution on the collaborative economy. It takes note of the importance of the collaborative economy in reference to the 2016 Commission Eurobarometer survey indicating that 17% of European consumers have used services provided by the collaborative economy. It also considers the following Commission communication on a European agenda for the collaborative economy: ‘A good starting point for promoting and regulating this sector effectively.27 While the Parliament agrees that the collaborative economy generates new and interesting entrepreneurial opportunities, jobs and growth and frequently plays an important role in making the economic system not only more efficient, but also socially and environmentally sustainable, allowing a better allocation of resources and assets that are otherwise underused, and thus contributing to the transition towards a circular economy, it acknowledges at the same time that the collaborative economy can have a significant impact on long-established regulated business models in many strategic sectors such as transport, accommodation, the restaurant industry, services, retail and finance; that it understands the challenges linked to having different legal standards for similar economic actors; and that it stresses the importance of ensuring a high level of consumer protection, of fully upholding workers’ rights and of ensuring tax compliance.28 In a specific section on the ‘Impact on labour markets and workers’ rights’, the Parliament notes that the collaborative economy is opening new opportunities to enter work for those who are unemployed, are currently far from the labour market or would otherwise be unable to participate in it, such as young people and marginalised groups, but under some circumstances this development can also lead to precarious situations. It therefore stresses the need for labour market flexibility on the one hand, and for economic and social security for workers on the other, in line with customs and traditions in the member States.29 The Parliament points out that all workers in the collaborative economy are either employed or self-employed based on the primacy of facts and must be classified accordingly. It also underlines the importance of ensuring the fundamental rights and adequate social security protection of the rising number of self-employed workers, including the right of collective bargaining and action, and also with regard to their compensation. On 26 April 2017, following a year-long preparatory phase30, the European Commission officially launched a European Pillar of Social Rights.31 Responding 27

European Parliament resolution of 15 June 2017 on European Agenda for the collaborative economy (2017/2003 (INI)), point I. 28 Ibid., paras 4–6. 29 Ibid., para. 37. 30 European Commission, Communication of 8 March 2016, launching a consultation on a European Pillar of Social Rights, COM (2016) 0127 final. 31 European Commission, Communication of 26 April 2017, establishing a European Pillar of Social Rights, COM(2017) 250 final.

122    Platform Economics to these initiatives, the Commission (2018) adopted the proposal for a council recommendation on access to social protection for workers and the self-employed. The proposal encourages EU countries to: (a) allow NSWs and the self-employed to adhere to social security schemes (closing formal coverage gaps); (b) take measures allowing them to build and take up adequate social benefits as members of a scheme (adequate effective coverage) and facilitating the transfer of social security benefits between schemes and (c) increase transparency regarding social security systems and rights. It covers social security schemes for unemployment, sickness and healthcare, maternity or paternity, accidents at work and occupational diseases, disability and old age. A report for the European Commission by Pesole et al. (2018, p. 53) argues that ‘even though a very large proportion of platform workers consider their work through platforms as a form of selfemployment (either primary or as side activity), a significant number do perceive themselves as employees of the platforms’. Given this context, the authors argue for lifelong training, linking entitlements to individuals rather than jobs, fostering mobility and mitigating the social cost of labour market adjustments, with an overall call for: a harmonisation of the conditions of platform workers towards those of regular employees. This includes access to benefits (see below), but also minimum wages and other issues. In particular, minimum wage policies can protect the income of a growing number of workers in low-wage jobs and their application to all workers should be considered. The reported intensity of working conditions also calls for a clarification of how health and safety at work regulations should be applied to platform workers. And there are also concerns about data protection issues, given the generalised use of platforms to monitor work performance. In Chapter 4, we offer an in-depth qualitative analysis of the rhetoric produced in digital intermediation platforms more broadly, with an ideological focus, by integrating further scholarship from digital activism and Internet studies, digital cultural industries and the political economy of platformisation to examine how discourses from sharing economy players might shape socioeconomic structures and relations of production, particularly in terms of mainstream and alternatives within the governance discourses of digital intermediation platforms.

Chapter 4

Ideological Production in Digital Intermediation Platforms* The organic nature of capitalist society is both an actuality and at the same time a socially necessary illusion. The illusion signifies that within this society laws can only be implemented as natural processes over people’s heads, while their validity arises from the form of the relations of production within which production takes place. (Adorno 2006, p. 118)

Introduction This chapter investigates practices and discourses of ‘sharing economy’ and alternative digital governance players from a theoretical perspective, combining critical analysis of platformisation and digital activism and political economy of communication, drawing upon in-depth interviews and practitioner event observation by examining an assortment of discourses. We find that there is an ideological spectrum, ranging from commons-oriented, peer-to-peer (P2P), decentralised, platformcooperativist and activist rhetoric, to the discourses of corporate and public organisations operating within the digital economy. Actors of digital intermediation platforms are, to a certain extent, conditioned by this ‘collaborative’ ideology, while hybrid (online and offline) organisational structures are effectively produced in direct relation to these specific ideological frames through experimental digital socio-technical systems. This development reshapes decisions in the digital

*A parallel much shorter form version of this chapter was published at the Journal Television and New Media with the title: Platform Ideologies: Ideological Production in Digital Intermediation Platforms and Structural Effectivity in the “Sharing Economy”, Athina Karatzogianni co-authored with Jacob Matthews. The authors would like to thank Hélène Falgayrettes and Ariadna Fernandez Planells for their generous support during fieldwork in Barcelona.

Platform Economics: Rhetoric and Reality in the “Sharing Economy” Digital Activism and Society, 123–150 Copyright © 2019 by Emerald Publishing Limited All rights of reproduction in any form reserved doi:10.1108/978-1-78743-809-520181005

124    Platform Economics political economy about business models, labour conditions, and how productive, commodity and money capital circulate in specific organisational structures irrespective of traditional varieties of capitalism. It also raises questions about the translation of distrust and suspicion of classical hierarchical formal centralised structures in favour of an ideological vision of informal, non-hierarchical organisations enabled by online communication platforms. This ideological production ranges from legitimising and reasserting a reformist, more humane capitalism to supposedly more radical visions of cooperative society, or to commons-oriented production, resisting privatisation through the recapturing of the public space as commons resources managed responsibly by communities with governmental participation. The political potential of platform activism points to the importance of ‘translating’ this new ideological spectrum into material relations of production. We propose that this involves questioning whether discourses that ‘cling’ to dominant social forms, demonstrate a higher effectivity than alternative ideological production aiming to surpass extant dominant relations of production. The rapid expansion of the digital economy is born out of varieties of capitalism across vast ranging national institutional frameworks – State–labour relations, re-regulations, privatisations, cross-class relations and coordination arrangements within diverse political systems (Hancké, Rhodes, & Thatcher, 2009). Yet, overall, the digital economy dances to the rhythm of two predatory forms of capitalist expansion, what Harman (2010) calls ‘zombie capitalism’, and Graham (2006) calls ‘hypercapitalism’. Zombie capitalism refers to Marx’s (1959) argument that once all capitalists introduce these techniques the value of the goods falls until it corresponds to the average amount of labour needed to produce them under the new techniques. The additional profit disappears – and if more means of production are used to get access to the new techniques, the rate of profit falls…. If some capitalists are to make an adequate profit it can only be at the expense of other capitalists who are driven out of business. The drive to accumulate leads inevitably to crises. And the greater the scale of past accumulation, the deeper the crises will be. (Harman, 2010, p. 72) Hypercapitalism refers to the trajectory of systemic capital extending the processes of commodification to include every aspect of existence: the social process in a knowledge economy, because of its focus on commodifying the products of language and thought, includes the entire network of activities and artefacts through which individuals and societies reproduce themselves at every level: materially, spiritually, socially, relationally intellectually, technologically. (Graham, 2006, p. 69) Therefore, the present working of digital economy under this double tempo of hyper-capitalism/zobie-capitalism does not provide credit to the notion of ‘postcapitalist society’, where citizens do not destroy but overcome capitalism (Drucker, 1993), and more for what Srnicek (2017) dubbed as a ‘platform capitalism’.

Ideological Production in Digital Intermediation Platforms    125 The proliferation of digital intermediation platforms occurs in diverse fields: cultural crowdfunding and crowdsourcing, content aggregation, advertising and marketing, on-line dating, car-pooling, ethical commerce and alternative finance, to name a few. Distribution, information and transaction occur in multi-sided markets, capturing positive externalities produced by the interactions of a multitude of players, including the tech giants who are generally not even be producing contents, goods or services of their own. As a result, the evolution labour has also been extensively theorised: the differences between audience labour (Smythe, 1977), cultural labour (Hesmondhalgh, 2010), digital labour (Cardon & Casili, 2015; Fuchs, 2014; Peters & Bulut, 2011; Scholz, 2013), algorithmic labour and platform labour (Andrejevic, 2009; Comor, 2010; Van Doorn, 2017). In short, the move from audience labour to digital labour to platform labour, forces the subjects to move from viewers viewing and consuming advertising, to users/prosumers engaging in “produsage” (term coined by Bruns, 2007) through playbour, consuming targeted advertising using them as products on social media sites, to workers selling their labour in the gig economy on platforms, whilst social protection becomes a thing of the past (see Gandini’s forthcoming for a detailed formulation on the evolution of the scholarship). Platforms are no longer merely cultural intermediaries (Matthews & Smith Maguire, 2014) but play on all tables: dead labour, intellectual labour, manual labour, audience, algorithmic and platform labour. Activities regarding the organisation of labour occurs on three levels (e.g. in cultural crowdfunding): within their own structures; filtering and editing contents, linking projects to external partners, often resorting to traditional forms of exploitation of cultural labour; and stimulating audience labour on external networks (Matthews, 2017). Crowdfunding and crowdsourcing platforms are producers of ideological discourses, busy promoting their short-term agendas and producing the illusion of modified relations of production and that of an inversion of the production cycle. In the spirit of the introduction of this book, which seeks to dissect the rhetoric of the ‘sharing economy’, we are connecting cognitive frames, social relations and organisational factors to elaborate on how the crisis of accumulation and hypercapitalist expansion affects socio-economic structures within the context of digital intermediation platforms. The role of the State and capital in backing non-profit platforms as providers of public service, making up what the State cannot provide and corporations can redeem as social responsibility tokens, is a red herring concealing the dismantling of local work forces into transnational online labour markets, which are either un- or under-regulated in terms of liability, taxation, insurance and social protection. In relation to the innovation versus social justice debate in platformisation politics, the promise of ‘objective governance’ through appeals to the magic of algorithms (search, coordination and transaction cost reduction) has so far failed to deliver increased employment and enhanced productivity, while new labour laws are radicalising workers across the globe, struggling against unsustainable capital accumulation relied upon unicorn notions of an environmentally conscious circular economy. In the next section, we present in brief the three main (sub)sets of literature that form the theoretical platform from which we launch our analysis: sharing economy, digital activism and critical political economy of digital governance and organisation specific to cultural intermediation platforms. This is followed by an

126    Platform Economics explanation of research methodology used, from which we conducted an ideological production analysis of platform actors, identifying core problematics, using 28 interviews, punctuated by considerations emerging from ethnographic observation and document analysis. This chapter concludes with core findings, limitations and future research potentialities stemming from this particular fieldwork study.

Integrated Theoretical Framework for In-depth Qualitative Ideological Production Analysis As explained in the previous chapters of this book, the paradox of the ‘sharing/collaborative economy’, defined simultaneously as part of capitalist production but also an alternative to it, is further complicated by the fact that much contemporary research into the political economy of platformisation relies on platforms’ own data and has been produced by platforms themselves or in dependent collaboration with due to the proprietary attitude platforms have about the data they collect. The owners of platforms rely on the future regulatory decisions, which are set to be fought in parliaments, in courts and on the streets. Despite the obvious differentiation between large privately own ‘gig economy’ platforms and smaller cooperativist style community-oriented platforms and the various in-between modalities, the management of internal and external labour is not a mere exercise in producing value, as it not only affects structural conditions cutting across industrial sectors but also produces particular ideological and cultural discourses, currently involving the recuperation of the commons and community as moral justification registers. For the purpose of analysing ideological production on platforms, we draw from three sets of literature: critical economy of platformisation, digital labour organisation and gig work, as well as digital activism scholarship. In this section, we set up the current environment, then we explain key debates in the sharing economy literature, audience/creative labour and digital activism scholarship, from which we launch three analytical sections, each examining three core ideological strands (dominant actors, commons and platform cooperativism), supported by fieldwork interview evidence. At the end of this section, we outline key entry points for our enquiry and the key research questions we sought to answer. In the first analytical section ‘It’s too ideological’, we probe deeper into both the ideological production and the aggressive strategies of intermediary players operating within what Kenney and Zysman (2016) identify as privately generated platform-based ‘ecosystems’, companies which fundamentally ‘are not delivering technology to their customers and clients – they use technology to deliver labour to them’ (Smith & Leberstein, 2015). In turn, Berg (2016) points out that platforms are not regulated by governments, but ‘this does not mean that they are not regulated, or that it is a free exchange of services between independent parties. Rather, the platforms regulate the market’. In fact, the platforms have a position ‘like that of the government’. The context of these platform wars according to Graham et al. (2017) is the following dystopia: The bargaining power of workers is undermined by the size and scope of the global market for labour; the anonymity that the digital medium affords is a double-edged sword, facilitating some types

Ideological Production in Digital Intermediation Platforms    127 of economic inclusion, but also allowing employers to discriminate at will; disintermediation is occurring in some instances, but the combination of the existence of a large pool of people willing to work for extremely low wages and the effects of the importance of rating and ranking systems, is also encouraging enterprising individuals to create highly mediated chains; and those mediated and opaque chains are, in turn, restricting the abilities of workers to upgrade within them. (p. 16) What’s more, we know from the iLabour index developed by Kässi and Lehdonvirta (2016) that despite the fact that ‘it is the information technology industry in each country that is currently making use of online labour’, ‘physical location of the contractors affects the contractors’ earnings, outcomes through the outside options in local labour markets faced by the contractors’ (Kassi et al., 2016, p. 6). Workers are also integrated within virtual production networks, and show that ‘while virtual product is embedded within networks and territories at various spatial scales, it is nevertheless, simultaneously marked by high levels of societal disembeddedness’ (Wood, Graham, & Lehdonvirta, 2016, p. 8). Lehdonvirta (2016, p. 14) points to the tension between placelessness and organisational identity, ‘where the means that are used to delocalise work – deskilling, codification, black boxing, algorithmic management – also undermine organisational identities’. As De Stefano (2016, p. 10) points out ‘the possibility of being easily terminated via a simple deactivation or exclusion from a platform or app may magnify the fear of retaliation that can be associated to non-standard forms of work, in particular temporary ones’. In a similar vein is the pessimism of Valenduc and Vendramin (2016, p. 41 cited in Degryse, 2016) who feel that it is hard to see a future for traditional working relationships in a world where digital platforms act as labour market intermediaries but with ‘possible lines of action are taking shape in the form of new trade union models, both on- and offline’. Benson, Sojourner, and Umyarov (2015, p. 23) view traditional labour unions and professional associations used for coordinating collective withdrawal of trade to discipline employers giving way to ‘the rise of new institutions that facilitate information sharing [and] may be taking up some of this role’. Regulation-wise, Masselli et al. (2016) suggest that while a sound European framework is crucial in cases of cross-border interest or single market relevance, cities should be allowed to promote initiatives that target the specific needs of local communities in strategic fields, such as sustainable mobility and tourism, and health and social services. Platforms can generate social capital and revive communities, but they can have skewed distributional effects detrimental to marginalised groups. This is, of course, partly linked to the co-optation of the ‘sharing economy’ by powerful platforms with strong public relations (PR) and corporate social responsibility (CSR) activities, stressing greener consumption and circular economy as well as wide-net social welfare gains. It is also reflected in polarisation between groups advocating self-regulation and those that argue for more top-down approaches.

128    Platform Economics Overall, as it has been argued in the book so far, the paths of the future of digital labour markets and their impact on society are influenced by the following approaches: (a) Changes will happen without requiring regulatory interventions through changes in behaviour and culture when the economy changes, head in sand optimism; (b) governments have to create new regulatory frameworks to correct disempowerment and unfair aspects stemming from fallouts in relation to how digital labour markets operate; (c) there is no need for social and cultural intervention from government because digital labour specialisation and migration occur ‘naturally’ with minimum government intervention and (d) algorithms, robots and AI embedded in digital labour markets will lead to severe crisis for traditional firms and work, which must be regulated by both governments and platforms with dialogue. It is critical to understand with some certainty whether digital labour markets improve production and efficiency, and if this is the case, then to assess the downsides of this improvement in terms of social and cultural impacts. Digital labour markets are less beneficial for developing countries and might drive down wages in developed ones. Moreover, there are collateral issues, such as the ‘superstar’ effect (where 10% of workers account for 80% of work completed), leading to job concentration and income inequality, little correlation between skills and earning levels, and with reputational ratings and references ruling the game, ethnic, class and gender-based discrimination is rampant. Accordingly, in the second (commons) and third (platform cooperativism) analytical sections, we investigate ideological production of alternatives in terms of digital labour resistance and new possible lines of action. Here, the ‘commons’, for instance, see Le Crosnier’s (2015) work on the ‘biens communs’ and Fuster Morell’s (2018) theorisation of Catalan ‘procomuns’, are relevant in the European context that we investigated and it is a too common ideological product in the actors we interviewed in Barcelona, Paris and Berlin. Besides the commons, there is considerable parallel influence from Scholz and Schneider’s (2017) efforts under the banner of ‘platform cooperativism’, an emerging network of cooperative developers, entrepreneurs, labour organisers and scholars developing an economic ‘ecosystem’ that seeks to align the ownership and governance of enterprises with the people whose lives are most affected by them. This represents a radical critique of the existing online economy, but it is also a field of experimentation for alternative forms of ownership design (Scholz and Schneider 2017). Scholz (2016a) looks to cooperative structures and the call for collective decision making, conflict resolution, consensus building and the managing of shares and funds in a transparent manner. He cites convincing tools that have emerged, such as Loomio, Backfeed, D-CENT and Consensys. In the summer of 2018, Scholtz’s worker solidarity attracted a million dollars funding from Google to develop a platform cooperativism kit. However, as all sharing economy scholars tend to point out, actions through these platforms tend to suffer from some of the typical problems of online activism, such as the effects of surveillance, oligopoly and corporatisation, reproduction of hierarchical and exclusionary systems and discourses, flash in the pan mobilisations and the issue of sustainability of movements, to name a few. Here we draw from the strands made in the scholarship on digital activism (Benkler,

Ideological Production in Digital Intermediation Platforms    129 2006; Bennett, 2004; Brevini, Hintz, & McCurdy, 2013; Castells, 2000; Chadwick, 2006; Diani & McAdam, 2003; Gerbaudo, 2014; Karatzogianni, 2006, 2015; McCaughey & Ayers, 2003; Milan, 2013; Rheingold, 1994; Taylor & Jordan, 2004; Trottier & Fuchs, 2014; Van de Donk, Loader, Nixon, & Rucht, 2004) and particularly in this analysis taking into account recent contributions made by Dolata (2017a, 2017b), Dolata and Schrape (2016) and Schrape (2017) and their concept of advanced technical sociality: ‘The institutionalisation of the collective can today no longer be represented as a purely social but only as a socio-technical process, understood as the systematic interweaving of social and technical organisation and structuring services’. In this respect, we argue that ideological production is emerging within socio-technical systems and is affecting and being affected directly by those very same. A study that supports this assertion is Ong and Cabañes’ (2017) research on the motivations and strategies of a well-organised and funded hierarchy of political operators in the Philippines, who maintain day jobs as advertising and PRs executives, computer programmers and political administrative staff but recruit a team of anonymous freelance digital influencers and fake account operators to seed core campaign messages in online spaces and create ‘illusions of engagement’ to inspire enthusiasm from real supporters. Among their motivation is a self-styled moral justification of ‘agent of social change’ against dominant structures. The third set of literature we draw from is broader digital political economy in the areas of Internet governance and oligopoly (Benkler, 2006; Castells, 2000; Jenkins, 2006; Lessig, 1999; Loader, 1998; Scholz 2013; Smyrnaios, 2017; Terranova, 2013 [2000]) and critical analyses of culture industries in relation to the ‘collaborative economy’ (Bouquillion & Matthews, 2010; Matthews, 2017; Pais, Gandini, & Arcidiacono, 2018; Nixon, 2014, 2017). A significant element of this theoretical subset is its concern for the questions of labour organisation and relations of production within the traditional culture industries, and that of their evolution at the current intersection of these industries with digital intermediation platforms. Important insights have been provided by the analyses of ‘digital labour’, whereas critiques of the ‘attention economy’ have illustrated the importance of generalised and automated data production and quantification of work (Moore, 2017). In relation to the innovation versus social justice debate in platformisation politics (see Dencik, Hintz, & Cableet al. 2016), the promise of ‘objective governance’ through appeals to the magic of algorithms (search, coordination and transaction cost reduction) has so far failed to deliver increased employment and enhanced productivity, while new labour laws are radicalising workers across the globe, struggling against unsustainable capital accumulation relied upon unicorn notions of an environmentally conscious circular economy. Several key aims of this perspective are underlined in the introductory words of Graham and Gandini’s (2017, p. 2): It draws attention to the kinds of creative collaboration afforded through digital platforms and networked publics. It considers how

130    Platform Economics these are incorporated into emergent market paradigms and investigates the complicated forms of subjectivity that develop as a consequence. But it also acknowledges historical continuities, not least in terms of the continued exploitation of Becker’s ‘support personnel’, but also the resulting conflicts, resistance and alternative models that attend the precarious nature of contemporary cultural work. Finally, it attempts to situate developments in the cultural sphere in broader social context and economic contexts, where (…) the idea of artistic collaboration has come to assume central importance. A significant element of this theoretical subset is its concern for the questions of labour organisation and relations of production within the traditional culture industries, and that of their evolution at the current intersection of these industries with digital intermediation platforms. Important insights have been provided by the analyses of ‘digital labour’, whereas critiques of the ‘attention economy’ have illustrated the importance of generalised and automated data production (Andrejevic, 2009; Comor, 2010; Fuchs, 2014; Hesmondhalgh, 2010; Peters & Bulut, 2011; Scholz, 2013). Nixon (2014) provides a novel analysis of ‘audience labour’, and the transformation of social communication into a process of capital circulation and accumulation. He asserts the continuity of a model where ‘capital’s ownership of the object of audience labour, culture, creates audience labour by creating a class relationship between those who own culture and those who do not’ (Nixon, 2014, p. 729). From this work, we question the importance of audience labour for capitalisation in culture and communication. Digital intermediation platforms indeed produce little or none of the digital culture over which they assume control, although that control is what allows them to extract surplus value from the consumption of that digital culture, that is, to exploit digital audience labour (Nixon, 2013). One therefore must consider whether the proliferation of web platforms effectively marks the disappearance of the culture industries, as we have ‘understood’ them for the past hundred odd years. This phenomenon appears to correlate with the proliferation of discourses and (often extremely mundane, semi-automatic) practices which escort new intermediation platforms in very diverse fields: Should these be seen as a formidable expansion of ideological production far beyond the frontiers of former culture industries (Matthews, 2017)? Further, we use Garnham’s (1979) work to understand the relation between the material and the ideological realms, in particular his three-level declension (a) material relations of production, (b) social forms of these relations of production (as in the capitalist ‘economic’ form of waged labour and (c) cultural forms of these relations of production. This type of ideological production analysis examines intermediation platforms as producers of material relations of production, social forms and cultural (ideological) forms. It interrogates how digital intermediation platforms are positioned in relation to (and feed into) the discourses of ‘commons’ and ‘collaborative economy’. Here, we are discussing Garnham’s (1979) hypothesis that the more autonomous a cultural form is with regard to social form and relations

Ideological Production in Digital Intermediation Platforms    131 of production themselves, the less effective it is (either for opposing them or for re-enforcing them). In other words, we are approaching the problem of low effectivity of many oppositional cultural products and discourses and stronger effectivity of discourses that ‘cling’ to the social forms of capitalist exploitation and the material relations of production that these are based upon. Drawing from these debates across diverse sets of scholarships provides us theoretically with the analytical tools to launch our enquiry. First, in respect of critical economy of platformisation scholarship, we wanted to know whether there is a common language among platform players with regard to ‘the commons’, ‘open’, ‘collaborative’ and so on, or an oscillation between different varieties/ iterations of capitalism with a ‘sharing’, ‘commons’, ‘cooperativism’ justification register. Second, in respect of digital cultural economies literature, we were eager to see whether they are expanding ideological production beyond former culture industries and whether this is superficial or substantial. Third, in respect of digital labour organisation and gig work, we examined what are the new cultural forms of relations of production the participants advocate; what relations of production have allowed this product/service that they have produced to exist; whether the participant’s answer simply served to legitimise their operations as ethical and/or politically radical or whether they were actually engaged in redesigning real labour processes, and in what ways. Fourth, in respect of digital activism scholarship, what kind of social relations do they legitimise, oppose or resist; terms of ideology (superstructural attributes) and structure (what are they potentially changing as actors in terms of economic value and cultural form)? Lastly, at the heart of our research questions, was the aim to understand how participants articulate these two realms (ideology and structure), and how these interact in the participants’ view of their work individually and within their organisation to verify whether Garnham’s (1979) hypothesis is confirmed, that is, the more autonomous a cultural form is with regard to the social form and the relations of production themselves, the less effective it is?

Methodology: In-depth Interviews, Participant Observation, Secondary Document Analysis To get more theoretical leverage, we draw empirical attention to the rhetorical foundations of the ‘sharing economy’ and the effect of ideological variants on the formation of diverse models, organisations and modes of production in the network economy by analysing the views of platform actors we interviewed in Barcelona, Paris and Berlin between November 2015 and February 2017. This in total included 25 trips between two researchers observing five international practitioner events (Procommuns, Transmediale, Ouishare, P2PValue and Cultura Viva), several protest events (Nuit Debut in Paris, Nit Dempeus in Barcelona and several anti-labour law protests in Paris) and the organisation of three expert workshops at the Open University, Barcelona (June 2016), Paris 8 University (April 2016) and the University of Leicester (December 2016). We interviewed 28 actors from varied institutional settings, from platform representatives (such as Uber, Airbnb and crowdfunding sites), sharing economy

132    Platform Economics watchdogs to platform cooperativists, public players, commons-oriented alternative governance groups as well as digital activists and artists. The study was a joint investigation, Foundations, Discourses and Limits of the Collaborative Economy: An Exploratory Research, bringing together and extending two projects: Karatzogianni’s ESRC project, ‘The Common Good: Ethics and Rights in Cybersecurity’ (project between University of Leicester and University of Hull) and Matthews’ (2017) research on ‘collaborative economy’ within the Collab research group at CEMTI (Paris 8 University). During data collection, we explained to the participants the purpose of the research and the interview process, and their right to withdraw at any time. The interviews were semi-structured and revolved around the following themes: understanding of their role in the organisation they operate in; their reflections about the sharing economy, digital governance and digital activism; their explanation of what forms of labour they engage in; their view on how their ideology, organisation and labour might differ to dominant and traditional organisations; their reflections on particular terms such as ‘sharing’ ‘collaborative economy’, ‘commons’ and ‘alternative digital platforms governance’; and their ideological inclinations and preferences regarding their work on digital intermediation platforms. The interviews were conducted in English and then transcribed professionally. The participants’ age ranged between 25 and 60 years and all had higher education qualifications. In more detail, we wanted to find out how interview participants talk about their own ideological orientations and their relation to the ideology(ies) of the Table 4:  List of Participants Participant

Nature of Participation

Date and Location of Interview

 1.

Participant 1

Smart city consultant

January 2016, Barcelona

 2.

Participant 2

Collaborative ecosystem January 2016, Barcelona actor

 3.

Participant 3

Tech-access activist

February 2016, Berlin

 4.

Participant 4

Digital artist/activist

February 2016, Berlin

 5.

Participant 5

Digital artist/activist

February 2016, Berlin

 6.

Participant 6

Architect/activist

February 2016, Berlin

 7.

Participant 7

Hacker/activist

February 2016, Berlin

 8.

Participant 8

Tech consultant

February 2016, Berlin

 9.

Participant 9

Securitisation expert

February 2016, Berlin

10.

Participant 10

Commons crowdfunding platform manager

March 2016, Barcelona

11.

Participant 11

Digital activism expert

April 2016, Paris

12.

Participant 12

Commons activist

April 2016, Paris

Ideological Production in Digital Intermediation Platforms    133 Table 4:  (Continued) Participant

Nature of Participation

Date and Location of Interview

13.

Participant 13

Open food business actor

April 2016, Paris

14.

Participant 14

NGOs actor

April 2016, Paris

15.

Participant 15

Tech developer

April 2016, Paris

16.

Participant 16

Tech developer

April 2016, Paris

17.

Participant 17

Digital game developer, documentary, activist

April 2016, Paris

18.

Participant 18

Movement activist media expertise

June 2016, Barcelona

19.

Participant 19

Movement activist international comm

June 2016, Barcelona

20.

Participant 20

Tech activist

June 2016, Barcelona

21.

Participant 21

Tech activist

June 2016, Barcelona

22.

Participant 22

Tech activist

June 2016, Barcelona

23.

Participant 23

Competition authority officer

November 2016, Barcelona

24.

Participant 24

Competition authority officer

November 2016, Barcelona

25.

Participant 25

Competition authority officer

November 2016, Barcelona

26.

Participant 26

Public policy representative, Uber

November 2016, teleconferencing

27.

Participant 27

Public policy representative, Airbnb

November 2016, Barcelona

28.

Participant 28

Platform cooperativism activist

February 2017, teleconferencing

socio-technical structures they operate in; in what relations of production do participants operate in their own initiatives/projects and how do these differ (if indeed they differ) with predominant capitalist modes; in what ways and to what extent do participants contribute to discourses of ‘sharing’/‘collaborative economy’, ‘commons’ and ‘alternative digital platforms’ governance’ and whether this is superficial or substantial. We wanted to understand whether the participant’s answer simply served to legitimise their operations as ethical and/or politically radical, or whether they were actually engaged in redesigning real labour processes, and in what ways. We also asked about what forms of labour the participant is engaging in and how they talk about their labour in terms of ideology (superstructural attributes) and structure

134    Platform Economics (what are they potentially changing as actors in terms of economic value and cultural form?). In this line of investigation, we examined how participants articulate these two realms (ideology and structure) and how these interact in the participants’ view of their work individually and within their organisation.

‘Sometimes It’s Too Ideological’: The Challenge of Collaborative Players to Steer the ‘Conversation’ We begin our analysis with the public policy spokesperson representing Uber in Spain (Participant 26, November 2016). After a short political career as a rightwing member of the Spanish parliament, he opted to join Uber, claiming that it is ‘probably the sexiest company right now in the world but also one of the most challenging ones’. He views his role as a public policy spokesperson as representing ‘what we contribute to society and to consumers how we can help cities change mobility in the twenty-first century’. Uber arrived in Spain in 2014 with a purely P2P model, which was challenged in Spanish courts. Uber operates in Madrid, working with professional drivers providing the technology service, but not in Barcelona, where the company instead launched a pilot project for delivery business. Pastor explained that part of his role is to translate to the media that ‘we are working with licences not P2P – we can provide more flexible and efficient way of doing things’. There is an issue of adapting the Uber model to Spanish law: We are working with professional drivers in Spain, as in most places in Europe. In Paris, London, in the UK in general, we’re working with professional drivers too. We’re trying to make the drivers with licences use our platform, we don’t have cars, we don’t have drivers, we just provide a technology. In the UK, they’re using a PHV licence, ‘private hire vehicle’, in France they’re using VTC and in Spain there are two types of licences for professional drivers: taxi licences and VTC licences. In Spain, there’s a short supply of these licences because there’s a law coming from 1998 that limited the VTC licences at only one for every 30 taxi licences so it’s an unequal balance. We’re now pressing in Madrid because there was a law between 2009 and 2016 that allowed people asking for more licences to get those, the court granted those licences, so the ratio I mentioned before, 1 for 30, isn’t applicable to Madrid. For an unknown reason, we noticed a higher demand for licences in Madrid than in Barcelona. For the time being, the number of licences in Barcelona is too low to launch a product with the minimum quality and standard. The regulation problem is a critical one, and at other points in the interview he was at pains to explain this and the frustration for a new ‘sharing economy’ player dealing with regulations that are not fit for purpose: The difference between London and Barcelona – and in London we have a huge business and a very successful product – is a light

Ideological Production in Digital Intermediation Platforms    135 regulation model: to own the licence, you only must meet certain requirements and then you are automatically aware that a licence isn’t a cab. Here you have a cab and given that the number of taxi licences hasn’t changed in the last 30 years, it’s not possible for the VTCs to increase even though there’s an obvious demand for this kind of service. He views the recent period as very unstable from a political perspective, particularly with regard to implementing policy changes in favour of Uber. With licence attribution being in the hands of local authorities, but the legal framework shaped on a national level, in Madrid, he sees radical political players (such as Barcelona mayor Ada Colau) as not being conducive to the start of what he calls ‘a conversation’. One example he presented of Uber’s supposed contribution to society is a recent episode of high pollution levels in Madrid, as he argued: We have cars in abundance in cities and we could provide more efficient services such as Uber Pool to get more people in fewer cars. It’d be very beneficial for riders who’d pay less for the same distance. It’d be beneficial for riders because the prices would go down, it’d be beneficial for drivers because they’d get more people in their car[s] and they’d make more money at once, it’d be beneficial for cities that would reduce the car use. When we asked about protests, labour resistance and media controversies in relation to Uber around Europe, but also globally, he pointed out: Whenever there’s a change, social or technological change, there is a resistance and this is quite understandable because some people can feel they’re kind of losing something with this change. We don’t think we are competing with the people offering the traditional services; we think we’re adding more efficiency to the way we use cars. There is significant political opposition and pressure in Barcelona against Uber and we press him about his thoughts on that: ‘Yes, you have city with a lot of demand for this service, a city that tries to position itself as an innovative place in Spain and Europe but in the same time, the regulation tries to close the possibilities for new services likes ours’. He declares that that pressure from incumbents is high and pressure has been exerted to cancel events where Uber was invited. When we questioned him on the strategies implemented to oppose this, in the media for instance, he answered: ‘Basically, it’s all about explaining what we can bring to society and show how we can do it, what kind of contribution we can provide. The only way to make sure something is going to change, is to get a lot of people into it’. The ‘conversation’ with political players and public authorities, as well as allegedly outdated regulations, are also burning issues for the Airbnb public policy

136    Platform Economics representative for the Iberic peninsula (Participant 27, November 2016). He has formerly worked as a civil servant for the government of Catalunya: What we have found challenging since we started is that Catalan regional regulations have been designed in a way that corresponds more to regular, not particularly progressive development, of old professional tourism regulations and which are applied to the new rule of the ‘prosumer’, this citizen who becomes both customer and producer. The old-fashioned approach to tourism and the electoral calendar have not been useful, he argues. The ‘conversation’ has become more complicated with tourism as a hot topic during elections. He claims: The conversation with the city officials in other cities is taking place in a longer perspective, in a more relaxed environment, where the policy makers can develop the agenda and work together, identifying the kind of users’ model experiences we want to promote together, whereas in Barcelona, from the very beginning, it has been very difficult from a purely political point of view. Definitely, these political balances are preventing innovation, by not allowing a reasonable relaxed playing field for policy makers, officials or ourselves. When we ask the Airbnb representative about the negative media coverage that Airbnb has received and where it may originate from ideologically, he pointed to pressure from both hotels and other private sector interests as well as Left-wing movements. He explains that Airbnb has had negative press coverage in Left-wing and Right-wing media; according to him ‘the hotels and the regional government of Cataluña, which is rather centre right wing, are also lobbying for the tourism policies’. At this point, we questioned his opinions regarding opposition from Left-wing political players, and in particular, representatives of the commons movement, and in that respect whether he sees himself as part of the collaborative economy: On one hand, it’s a matter of ideological departure point, or philosophical starting point, and on the other hand, the effect that externalities may have in the debate: institutional arrangements, or a new political agenda, a new political calendar, but also the impact or negative impact tourism can have on the city and on its different districts. This influences a lot the conversation, but we’ve also noticed that, at times, the Left-wing movements in Barcelona don’t reflect on the positive impact that the sharing economy has for the little guy, for families, for middle-class people who really have an opportunity to get an extra for themselves. Sometimes it’s just too ideological. Here, in Barcelona, unless the sharing

Ideological Production in Digital Intermediation Platforms    137 economy is based on the pro-commons movement or the cooperative movement, it doesn’t exist; we close the door, we don’t want to listen anything about that and it becomes so ideological as well and so reluctant to innovation in a broader perspective. The Airbnb representative understands the dominant players in Barcelona as three big groups, which are as follows: Firstly, the main telecom operators and large corporate groups; secondly, a powerful start-up community and thirdly, movements promoted by the city hall: cooperatives, social economy and commons players. He believes that city funding of these new players is more relevant than the funding of the start-up community: ‘They have their own lives, their own apps: they don’t depend at all on public funding’. In the discussion about ‘for-profit versus non-profit’ and ideological production pertaining to this distinction, the Airbnb representative makes a point regarding ‘businesses versus not businesses; the only non-profit organisations are NGO, I mean any business, whatever the sector, they go for profit and that’s an important point. Even the cooperative companies go for profits and they have to be careful on making money’. We come back to this assertion several times during the interview. He insists that he is ‘comfortable’ with the city hall’s policy of supporting commons-based initiatives, and ‘super happy they are promoting this kind of economy’, with which he considers Airbnb is fully compatible: We believe that the social [and] cooperative economy is fully compatible with the sharing economy movement, with our users’ community. We want to get as close as possible with them; we’d really like to work with them and actually we are starting to do it, organically. The only thing I wanted to point out before is, when they say ‘you are capitalists and you only care about money’, well we have to keep in mind that everybody in this field is a business and go for profit, we should not forget the corporate social responsibility and they should not forget that they are businesses and they’re part of our world, we shouldn’t be splitting those two worlds. He feels that Airbnb is really a novel and small player in the ‘conversation’ around tourism. He explains that the city of Barcelona has been talking about tourism for the last 20 years, and the issue has become more sensitive over the last 10: We are really new in this conversation and when it comes to power, well if you compare the power that the sharing economy has to [that of] the city hall, the hotels associations, the regional government, the newspapers, we are really small. However, we are really strong in terms of how the consumer is favouring this new model.

138    Platform Economics When asked about Airbnb’s future development prospects, he claims: Airbnb will most probably evolve into a world of experiences. That is the future, that’s a thing we have learned over those years, it’s not about accommodation, it’s about finding a new way to connect the people and make them belong anywhere. With the regulation issue for newcomers in the ‘sharing economy’ in Barcelona identified as a core discussion point, we interviewed the Catalan Competition Authority to investigate further their approaches and recommendations in relation to these players (Participants 23, 24 and 25, November 2016). This is a publicly funded anti-trust body covering two fields: competition law and promotion of competition. The authority studies firms in terms of their undertakings and examines regulations from the municipality, regional and national governments, but are only responsible for Catalonia (the Spanish competition authority being in charge of broader cases and issues and also answering to the EU). The ‘sharing economy’ cases they have engaged with regarded both legal aspects and the promotion of free trade. We interviewed three officials, and their criticism echoes the concerns raised by both Uber and Airbnb representatives: previous regulations aren’t fit for these new players. An officer we interviewed from this unit suggests that a lot of innovation is needed to change regulations (Participant 23). We ask how ideological leanings might influence the competition authority, to which their replies are: ‘Our objective is to guarantee free competition; we are in favour of free economic activity (…) but we recognise externalities and we understand regulation has a role’. In relation to tourism, they feel that Barcelona has to manage and solve this problem: ‘We are trying to help them by proposing measures that they could apply to balance the externalities and to guarantee economic activity’. Here we asked about a specific episode when Barcelona mayor Ada Colau used her powers to temporarily restrict the tourism market and review its development (the municipality ceased delivering new licences for rooms within the city centre and for non-sustainable accommodation in the periphery). At this time, the Competition Authority published a report, making recommendations based on transferable licences and openly criticising the move: ‘By not giving any more licences for four years, you are not allowing anyone entering the market, so in a way authorisation itself becomes an asset’ (Participant 25). When asked whether they are frustrated with the local government, the director general replied, most diplomatically: We are waiting; there is no frustration; we understand things go slowly. There is a working commission for the sharing economy; they are analysing how the regulation should be modified, we are happy about that. It could work faster but ok. (Participant 24) When we pushed to understand more of the ideological tenets of their organisation (i.e. whether they see themselves as politically neutral, as a public service etc.), the

Ideological Production in Digital Intermediation Platforms    139 response was: ‘The more companies we have on the market, the better it is, because the prices are lower, we have a better quality, more innovation’. We put forward to them that if their default position is free competition, this is already an ideological position, to which we got the astounding answer: ‘Yes’ (Participant 23). These elements point to the effort deployed by ‘sharing economy’ players in order to be ‘part of the conversation’, as it was put by Participant 2, a collaborative economy consultant we interviewed in Barcelona, in January 2016. Indeed, in both emblematic fields of personal transportation and accommodation, the question of opening competition to digital intermediation platforms operating outside existing regulatory frameworks, has become what Airbnb representative named ‘a hot topic in the political agenda’, which supposedly allowed for wide coalitions against the ‘collaborative’ players and therefore ‘made the conversation more complicated’. The Uber representative concludes by saying: I think that now the situation is warmer and we can start to have a conversation (…). So far it has been very difficult to have any kind of conversation because when you’re holding elections this is not possible to have it. The notion of a ‘complicated conversation’ (or an ‘impossible conversation’ due to elections) is worth noting; indeed, it does refer specifically to the political context of Barcelona (where support for a radical Left coalition was strong enough to lead to the election of Ada Colau as mayor). One might well ask whether this is also, more significantly, a way of referring to public debates which sporadically and temporarily escape the streamlining and steering efforts of corporate communication agents. In dominant sharing economy actors and governmental regulation, ideological production draws heavily from a neo-liberal position. In the next section, we investigate the ‘commons’ ideological spectrum as a competing ideological production.

Against, with and Beyond State and Capital: Commons Discourses, Multifarious and Paradoxical The notion of commons and commons-oriented production was ‘spontaneously’ present in over two-thirds of the interviews we conducted, and we focused on this specific element of ideological in the discourses of all our interviewees. The first illustration of this comes from the interview we had with a representative of Goteo, a Barcelona-based crowdfunding platform (which happens to be promoted by the municipality), at ‘Cultura Viva’, an event we attended in March 2016. Goteo claims to be dedicated to providing funding for projects that are both commons-oriented and socially inclusive and sustainable. We interviewed platform manager for Goteo (Participant 10, March 2016), who explained this key condition for obtaining funding via their platform: To get the funding you have to be committed to open up and commonise your outputs for society to use, where you open up

140    Platform Economics your outputs and offer them for the community to develop further create derivative work from [them]. The idea is that if you commonise your sources, you are preventing privatisation because you are making them for the community to use. (…) You have to have a tool that we could implement to produce change in the democratic field to be more democratic, more transparent more participatory-oriented. Their online platform is represented as such a tool, promoting the values of the commons by supporting organisations and individuals who develop projects for the benefit of specific communities. We interviewed a Goteo user (Participant 17, April 2016), a digital game artist/activist who raised funds for the production of a documentary film illustrating the implementation of wireless mesh networks in rural communities in northern Greece, and how this also contributed to the development of more or less autonomous production processes (notably in the fields of agriculture and crafts). He affirms: ‘We thought that it was a good occasion to launch not only that crowdfunding [campaign] for the documentary but (…) in general (…) the idea of crowdfunding for Greek social movements’. It is worth noting that he considers that his own experience of crowdfunding a documentary using Goteo can be transferred to the entirety of ‘Greek social movements’, and that web-platform-based collection of funds (and labour) represents a remedy against the exhaustion of social and political groups having previously relied on traditional fund-raising via physical donation requests and organisation of events. He mentions the capital control measures instigated by the European Central Bank and the International Monetary Fund in June 2015, and points to a paradox: ‘Although, the Greeks couldn’t use their debit cards directly, they could use Paypal’. Later, he declares: In general, people that are doing the crowdfunding organizations, campaigns and platforms, are, not all of them, but there is a spirit of what is called techno-optimism. This techno-optimist spirit means that, with the right tools and the right knowledge on networking connections, we can (…) solve problems. We realised that if the international financial elites want to act on a country, on a network, on a system and take decisions on the financial level, then any kind of platform reaches its limits. Indeed, before this the Goteo team hadn’t faced the problem of capital controls. Regarding this, he argues: ‘There is need of political organisation to put pressure, as nothing can go on if there is not a political body that functions off the cloud, and doesn’t depend on the cloud, knowing how to apply pressure to power structures’. This experienced activist and crowdfunding user goes on to express what he considers as one of the key problems with crowdfunding: Collaborative economy projects (…) are more and more re-appropriated by private institutions not only as methods and as crowds,

Ideological Production in Digital Intermediation Platforms    141 as money finally, but also as linguistic, semantic structures. For instance, (…) three days ago, I received an email from a big, private cultural organisation in Athens [which is] very aggressive, aggressive to public space. I mean that they are doing a crowdfunding campaign to finance one of their projects. And, they use the same language, the same vocabulary that we used for our crowdfunding campaign. It could be even ‘copy-paste’. I don’t mean by that they copy-paste me or our campaign. But they copy-paste the movement the same way that Syriza in government copy-pasted the slogans [used on] Syntagma Square five years ago. (Partcipant 17) What is implied here (given that the interviewee has previously stated that crowdfunding began as a ‘grassroots’ movement within independent music scenes) is once again a typical them/us dichotomy; here with capitalist ‘creative industry’ players mimicking the language of the opposition and corrupting the formerly pure ‘collaborative’ tools in the same way that Alexis Tsipras’s government hijacked opposition movements to seize power and serve its capitalist masters. Meanwhile, the semantics of ‘collaboration’ may represent a common language from the word go – and indeed, the ‘very aggressive’ private cultural organisation he mentions can be just as legitimate and indeed as efficient as his own organisation when it comes to implementing this type of strategy to collect financial resources. A commons activist we interviewed in Paris in April 2016 (Participant 12), during the Nuit Debout mobilisations on Place de République, understands the commons not only as collective action but as resource: You have to act for the commons but the common pool resource might be something you build, but it can also be something that is global and universal, but you have to transform it from public to common. Something that nobody owns [is] universal; it becomes common when people try to come together to defend it. He seizes the example of Parisian mayor Anne Hidalgo’s condemnation of Nuit Debout allegedly ‘privatising the public space’ by their occupation of Place de la République: ‘In fact, they are not privatising, they are transforming the public domain, the public space, into a commons, by their activity in the commons’. He goes on to ask: ‘How can we have a new partnership between the State and the commons?’ In the section that follows, we conduct an ideological production analysis in the analytical set of platform cooperativism. Participant 19 (Barcelona, June 2016), who was an international communication activist for the 15M movement in Barcelona, comments on the confusion between the private, the public and the commons: The commons is a source. It’s the knowledge that your grandmother, grandfather has pitched to you on the way to do things … Because (…) there is a heritage, there is something that we have

142    Platform Economics to keep, to preserve. And I think the political idea of commons is misunderstanding the public and the commons. I’m not supporting a political party to support the commons, but to leave the commons alone and to recuperate the public. And to stop the privatization, this is bad. (…) I’ve been doing lectures in Eastern Europe about commons, and most of the main questions are that people don’t understand why private is bad. I mean, they are in the 90s. They are in the wave of everything should be privatized. According to him, the idea of commons being opposed to the public remains inaccessible for wide sections of society; he and others (such as urban commons activist Participant 6), interviewed at the ‘Transmediale’ event in Berlin, in February 2016, openly suggest that such a confusion is actually managed and administered to negatively impact decision making about the use, future and maintenance of common resources. Let’s now briefly return to Participant 12, the commons activist, who suggests that crowdfunded projects on Goteo are one of many incarnations of the commons (as they allow distribution of surplus wealth to wider communities) and asserts that there is a broad ideological spectrum of the commons. Indeed, he spontaneously declares: ‘We need to maintain this idea of the spectrum, not to be too rigid’. It’s worth questioning the correlation between this affirmation of the need to maintain this idea of a wide spectrum and the vigorous refusal of such a continuity that one finds in the discourse of Participant 17 (or that Participant 12 perceives, for instance, in Richard Stallman’s ‘dogmatic’ defence of free software). Whereas Participant 17’s commons/collaborative players are seen to be in a minority position, fending off the old, proprietary vampires of Capital, Participant 12’s players are perceived as slowly but surely winning the battle against these same old forces, subverting and integrating them into the ‘spectrum’. One can also find echoes of this dichotomy when comparing the assertions of two prominent representatives of Ouishare interviewed for this investigation (which we will return to in the following section); one advocating integration of the whole spectrum of ‘collaborative’ production, from large corporate outfits to individual hacktivists, the other claiming that ‘the corporate follow us or lag us (…) but they will not be the leaders in this game’. One cannot but seize this opportunity to pose the hypothesis that the myriad of commons-stamped corporate agents and allegedly pure ‘collaborativists’ are objectively united by material relations of production that give little currency to these seemingly conflicting ideological forms. We met Participant 13 of the Open Food Network at Ouishare event in Paris in April 2016. She views the commons primarily from the point of view of infrastructure: How can we collaborate and co-manage the infrastructure that we need for our lives? So, for example, Open Food Network is what we call a commons in the sense that it’s a web application, an infrastructure which is open source – so the code of this platform is free, is shared openly, and anyone can take it and build his

Ideological Production in Digital Intermediation Platforms    143 own project of it. So, this is a commons, and we build around this open source code – we build collaborative, democratic governance entities. She talks of local entities in nine countries, which have chosen forms of cooperation which are ‘not capitalistic structures; they are governed by people’. She explains that the political mission behind Open Food Network is ‘to decentralise the food system, to enable and empower people and communities so they can build their own distribution channels directly connected to the producer’. In her view, this empowers producers and relieves them of the pressure and power that central players exert today within the food chain. When asked about the business model, she declares: It’s a non-profit project, but of course we have some costs that we need to pay for. But, here again, every local community builds its own business model. For example, in Australia, we take 2% commission on the sales, capped to $50. In France, we are thinking about taking a 2% commission but leave the option to the hub, to the food hub, to compete in a different way if they prefer. So that also we can empower hubs to innovate and build innovative business models. For example, one of the first beta testers in France is experimenting with volunteer contributions. So, they ask their clients to give to the platform. We’ll see what happens, but we like to make it possible for people to experiment. (Participant 13) Given the platform’s present scale of development, it seems obvious that allowing each local community to test varied business models is the only possible option. Asked whether her ideological orientations concord with the actual realisations that she is contributing to, she answers that although she is passionate about the project, it has been ‘hard to pay for my living expenses in the first years’ (Participant 13). She evokes two vague revenue sources: payment in the form of commissions, or building hubs ‘with different options to earn money’, and adds: ‘I really do believe that we are transforming the system. We need to get our hands dirty [but] I believe in decentralised systems that empower us to create alternatives’ (ibid.) This type of reasoning for allowing people to ‘experiment’ with the commons can also be found in the discourse of Participant 1: I think [that] the first step is to provide opportunities to equal access to the resources. Also, to have discussions around the governance of the commons so that those who are active contributors can also regulate how the commons are being used. Because you could perfectly use a common resource for something that is very negative for society, and that creates tensions. But it is my understanding that those who are involved, and will suffer the consequences, are the ones who need to negotiate this, rather than me, coming in as an outsider to try to decide what is common good and what is not.

144    Platform Economics Throughout these varied attempts to frame the notion of commons, the nurturing of ‘alternative’ systems of governance and/or resource allocation appears in the guise of attempts to either redesign or replace existing corporate or institutional intermediaries (however remote such a prospect might be). Beyond the ‘variable geometrics’ of the commons discourse, one might well ask what fundamental innovations are introduced, in terms of business models and administration procedures, within the various organisations and structures mentioned; i.e. are power relations, and relations of production, effectively modified, or do these processes simply amount to substituting groups of existing rentiers for renewed or upgraded ones? From this point of view, it’s interesting to note Participant 1’s response to a question we posed, urging her to define the commons-based governance models she advocates in her consultancy work: I don’t think it is fruitful to compare what we have, or what we put forward, with what would be the ideal set up, because we’re always going to fall short, so it’s unfair. What we’re trying to do is upgrade reality one step further.

And in the Name of Platform Co-operativism: ‘We’re Interested in Exploring the Whole Spectrum of Options’ We interviewed Participant 2, Ouishare representative, in Barcelona, in January 2016. Evoking the avatars of Barcelona’s ‘collaborative’ scene, he mentioned a specific group of Airbnb hosts that were planning to split from the mainstream lodging platform: They are thinking by themselves on creating a cooperative in order to do the invoicing in a legal way and so on, so you also see that at the end the peers can coordinate themselves, (…) you see this counter power because well-organised peers can have a similar power that the platform can have because a platform without the peers is nothing. Almost half of our interviewees ‘spontaneously’ spoke of online cooperatives and digital cooperativism. One prime example is Participant 28, who created his first online cooperative in 2003, and then went on to create a platform cooperative, which claims to be an ‘open global cooperative that organises itself through the Internet outside the boundaries and controls of Nation-States’. Moreover, it ‘aims to issue an alternative global economic system based on cooperation, ethics, solidarity and north–south redistribution and justice in economic relations’. When asked to define ‘cooperative values’, Participant 28 stated: Solidarity, mutual support, openness, to include new people and influence them to be consensual, participative, and so there are many user circles connected to movements. It’s also connected to open source, free hardware, all the digital movements related to the

Ideological Production in Digital Intermediation Platforms    145 commons. It’s about putting together many values to create something really equal, really fair and open and able to not just solve the problem but consider the whole thing as an inter-connected ecosystem (…). From my point of view, (…) it’s just an application of the traditional cooperatives but becoming digital and getting the capacities for people to cooperate on the platform. From my point of view, the platform, the digital spaces are more and more important, but for me it’s not enough because just a handful of cooperatives cannot fight in a capitalist society, so I think this platform should be part of an ecosystem in a very interconnected way. Recognising both the shortcomings of the online cooperativist movement and the immensity of the task lying ahead of him, he nonetheless suggested that the network he had set up was not simply about solving issues related to democratic participation and ownership but more fundamentally to the building of a new economy in a post-capitalist society. To understand this optimism, one should, of course, bear in mind that Catalonia, where a significant number of our interviews took place, has been historically marked by cooperativism in its anarchist and libertarian forms, ever since the second half of the 19th century and, in particular, during the Spanish revolution of the late 1930s. In this respect, it was interesting to observe the somewhat condescending appreciation of what Participant 2, for one, dubbed the ‘traditional’ cooperative movement, whose presence is strong within the radical left-wing coalition currently ruling Barcelona: Cooperativism has been very strong in this region for many, many decades, but in a very traditional form. These people are still attached to this very traditional form of low tech paper-based big meetings with big consensus and so on, and they are now a little bit in conflict with the technology. Nonetheless, this participant stated that part of his ‘mission’ was to reconcile what he claims to be two currents of cooperativism: Each of the groups can learn from the other. So the capitalists can learn how to have a better governance and better value distribution from cooperatives, and cooperatives can learn from the capitalists how to scale and how to have impact. Hence, ‘when I go to a cooperative movement I’m the capitalist. When I’m on the OuiShare movement I’m a little bit the cooperativist’. However, he did clearly stress: ‘I would not confuse cooperativism, which is a type of internal governance, with the production of common goods. It can go together, but it usually it doesn’t’. This affirmation is important from the point of view of numerous self-styled cooperativist players’ commitment to actually redesigning relations of production (either in a pro-capitalist form, increasing labour precariousness, or in an anti-capitalist form, aiming at abolishing exploitation).

146    Platform Economics This point was also interestingly summed up by the Spanish representative of Airbnb, when he declared that ‘everybody in this field is a business and goes for profit. We should not forget the CSR and they should not forget that they are businesses and they’re part of our world, we shouldn’t be splitting those two worlds’ (Participant 27). Is it not also what Participant 28 was implicitly advocating as he stressed the need ‘to build a new society based on cooperative values’? At a particular point in his speech, he resorted to the use of the Spanish language – ‘by putting together many values to create something really equal, really fair and open and able to not just solve the problem but consider the whole thing as an inter-connected ecosystem that really cambia la forma del capitalismo’. Indeed, this translates into English as ‘changing the form of capitalism’, which appears to refer precisely to what these various agents are talking about, i.e. reforming the cultural (and to a lesser extent the social) forms of capitalism. The participant is hopeful that the present working of the digital economy under the hyper-capitalism/ zombie-capitalism double tempo in our ‘Introduction’ Section will eventually provide for Drucker’s (1993) prediction of ‘post-capitalist society’, where citizens do not destroy but overcome capitalism. Here one also observes a typically post-modernist inversion between the ideal and linguistic, on the one hand, and the material, on the other: ‘putting together many values to create something’. We find the same process at work in the discourse of Uber representative: ‘This is what we do, we’re trying to explain how technology can help the cities to improve the situation’ (Participant 26). Again with Participant 1, who talks of ‘how technology transforms different aspects of society and economy and so on’. This coherent (or possibly false) contradiction is also at the heart of Ouishare Participant 15’s ‘mission statement’: What I wanted to do is to create an evolutionary environment where people could try. That’s my underlying core belief here. I also believe that we can actually create organisations that are more humane, more faithful to people that are engaged in them, and can also be profitable. Elements from the discourses of other interviewees are significant at this point. For instance, Airbnb rep argues: Yes, Airbnb will most probably evolve into the world of experiences. That is the future, that’s a thing we have learned over those years, it’s not about accommodation, it’s about finding a new way to connect the people and make them belong anywhere, so the next step will follow. (Participant 27) Uber rep similarly claims: ‘Basically, it’s all about explaining what we can bring to society and show how we can do it, what kind of contribution we can provide. The only way to make sure something is going to change, is to get a lot of people into it’ (Participant 26). Here we observe obvious attempts to redesign real labour processes: ‘connecting’ producers (and notably producers of the value platforms

Ideological Production in Digital Intermediation Platforms    147 capture rent from) and spreading effective discourse/experience in order to mobilise – one could say ‘agitate’ – the various groups of atomised service and goods producing and consuming individuals. In this respect, platform managers are not only in charge of capital/labour intermediation, they appear to be fully concerned with ideological agitation (‘connect the people’; ‘get a lot of people into it’; ‘make them belong anywhere’), an aim which is equally implicit in the declarations of the participant of Open Food Network (Participant 13), whose discourse suggests that users are striving to ‘change the world’, despite the fact that the platform objectively provides free labour for production processes that are ultimately intertwined with staunchly capitalist players. By the same token, one can but question the deeper motives of the seemingly gratuitous experimentation of Ouishare representative (Participant 2): Ouishare is super wide. So in, for example, the event that we have in Paris (…) in the same room, actually a large tent, we had the director for Airbnb for Europe (…) and a guy who lives without passport and without money; he’s a hacker and he joins projects, and Ouishare is interested in all that spectrum, all this empowerment, all this capability. (…) As Ouishare, we’re interested in exploring the whole spectrum of options. Then we’ll see what the result is in terms of impact and in terms of people’s empowerment. This wide-spectrum outlook clearly echoes the vision put forward by commons activist, Participant 12. However, the frame within which this ‘exploration’ takes place was, however, clearly defined earlier in the interview: optimised ‘governance’ and ‘value distribution’, capitalistic ‘scaling’ and ‘impact’. Moreover, it’s interesting to note how Participant 2’s discourse wholly replicates the rhetoric of web guru Tim O’Reilly. Indeed, Participant 2 states: The change is that you don’t capture the value like traditional capitalistic approaches [where] I try to capture value and make it private for me. (…) The big change is when this peer to peer production goes into the commons and this is a mindset that more and more people understand, that this is the more efficient way to share wealth. Instead of capturing the wealth only for me, how can I have this shared capital in the commons (…) for everybody, so that I will benefit from because it will be at my hand, but it will be at everybody else’s hand to use? For the record, O’Reilly’s own words, recorded at the San Francisco Web2.0 Expo in May 2010: Everyone is stronger than anyone individually and so it’s really important that we start looking at (…) the model of the Internet: small pieces loosely joint, everybody co-operating. My advice to all the players is to think about O’Reilly’s model, which is to create

148    Platform Economics more value than you capture. If you ask: ‘Am I creating more value than I capture?’ then you can start to quantify and you can say: ‘Yes I’m building value for users (…).’ (…) It’s paradoxical that when you create more value than you capture, you do create value for yourself. So, I would just urge all of you, as you build you apps, to think about being part of this co-operative Internet Operating System, not one where you try to own it all, but one where you gain strength by working together. It’s worth pointing out, if at all necessary, that such discourse is by no means innovative, nor does it depict any groundbreaking shift from the point of view of capitalist economics. The development of widely available industrial credit and a secure, global and privately owned banking system in the mid-ninetenth century can indeed be analysed as a similar form of capitalist resource pooling, in the same way that capitalist ‘usage’ of the State (quintessential for the securing of that very banking system) has also historically been a means of limiting risk for economic agents by effectively placing, as Participant 2 argues in somewhat misconstrued words, ‘shared capital in the commons’ and ‘[gaining] strength by working together’, similar to O’Reilly’s claims. What is, however, interesting here, is that for platform managers, ‘collaborative’ and self-appointed commons agitators, the diminishing roles of the State and of the traditional banking sector – although still both vital – appear to be occulted behind a shifting (if not simply shifty) representation of an autonomous crowdsourced and crowdfunded ‘ecosystem’ or that of an autonomous intermediation and governance system which might somehow apply to these players. It’s important to recall at this point the growing interest of existing financial and new fintech players (such as Orange or Telefonica) for what are effectively important new (or revamped) markets. In this respect, it’s worth considering the vague notions that Participant 2 deploys with regard to the funding of the Ouishare organisation that he is part of. When asked whether any funding came from the State, he replied: ‘No, it’s a private bank, a foundation of a bank in this specific case.’ In fact, Ouishare doesn’t appear to have been funded by any foundations belonging to banks but by the French insurance company’s ‘MAIF Avenir’ investment fund.

Conclusion The drive towards platformisation has gained significant impetus over the past 10 years. It does, however, remain a contradictory process, giving rise to significant resistance from both manual and intellectual labour, however poorly organised this remains as yet. As an opening remark, we acknowledge that although our study provides some insight, it doesn’t allow to fully validate Garnham’s (1979) hypothesis of a higher effectivity of less autonomous ideological forms. It does, however, offer support for this proposition and expands it, in at least four respects. First, we note the importance of ideological production for the players we interviewed. One can argue that this is their main activity as well as setting up

Ideological Production in Digital Intermediation Platforms    149 and running instruments for transaction and organisation of labour. One key element we find in all discourses is the imprecision and confusion of the ideological forms produced and in particular the forms (models and terms) used to describe relations of production. Simultaneously, all these platforms are at least partly dependent on commodity exchange; labour remains commodified, and none of our interviewees proposed any form of coherent plan to effectively transform relations of production. In fact, the ideological interchangeability displayed by these actors has an objective basis in material production, and that we can see from our interviews that they are in a position of relative dominance in comparison to the wider mass of network and platform users, and in particular to manual labourers whose activity is organised via these ‘tools’. Dolata (2017, p. 19) interestingly writes: The activists and participants of this type of movement are recruited from the pool of well-educated, dissatisfied and onlinesavvy young people of the urban middle class. Their self-understanding is characterized by a deep skepticism of the classic forms of organizing and the propagation of informal, non-hierarchical and non-ideological structures. Indeed, the dissatisfaction with ‘what they did before’ is palpable in many interviews, yet one might wish to critically question the assertion that such players are ‘well-educated’ considering that contemporary university’s function ‘as the training ground for cognitive capitalism’s immaterial labourers’ (Dowling, 2011, p. 195). Furthermore, this proposal could be improved by substituting ‘deep skepticism’ for ‘deep ignorance’ (of the classic forms of organising), and replacing ‘the propagation of informal, non-hierarchical and non-ideological structures’ by ‘the propagation of an ideological vision of informal, non-hierarchical structures’. Second, these players are also heavily involved in setting up socio-technical apparatus which is both what they talk about, what they ‘agitate’ for and what allows them to capture rent – however scarce – from processes of exploitation of labour. They are, from a material point of view, dependent on these apparatus/ platforms in order to survive in their current condition. Setting up, running platforms and spending a large proportion of one’s labour time in agitation is paramount for this individual survival, but it serves a goal which is much wider than simply allowing either individual or even that of wider social groups (‘commonsoriented digital activists’; ‘the collaborative ecosystem’): our hypothesis is that these players are in a sense inadvertedly (Marx’s, 1959) own term ‘involuntary promotion’) serving what Nixon (2017) calls ‘communicative capital’. Nixon (2017) talks of a class relationship, and therefore of class struggle between capitalists (‘communicative’ or belonging to other factions) and labour (whether ‘audience’ labour or traditional forms of labour, cultural, digital, platform or otherwise, see previous discussion in literature review). Nevertheless, Nixon’s (2017) sharp antagonism misses the intermediary nature of these actors which adhere more to the Fischer’s (1996, p. 81 cites Marx (1959, Capital, Vol. III, pp. 862–863) explanation of Marx’s (1959) ‘middle and

150    Platform Economics intermediate strata’, obliterating lines of demarcation, or a standstill type of crystallisation. It must be stressed that these players are not ‘communicative’ or other types of capitalists, even if to various and certain degrees they objectively advance the agenda of these last, pointing towards a certain degree of interchangeability. Adorno (2005 [1951], p. 128) correctly analysed this evolution, almost 70 years ago: The most powerful person is he who is able to do least himself and burden others most with the things for which he lends his name and pockets the credit. This seems like collectivism, yet amounts only to a feeling of superiority, of exemption from work by the power to control others. In material production, admittedly, interchangeability has an objective basis. The quantification of work processes tends to diminish the difference between the duties of managing director and petrol-pump attendant. It is a wretched ideology which postulates that more intelligence, experience and training is needed to run a trust under present conditions than to read a pressure-gauge. Here, we must question precisely, who these players are, from a class perspective? Here we suggest using the theoretical work of Alain Bihr (1989) who talks of a third, intermediary class ‘between’ capital and labour, or between capitalists and the working class, which he calls ‘capitalist encadrement class’ and mobilising both classical Marxist theory and Bourdieu’s (1984, 1986) analytical framework, set out to consider class according to the following four linked criteria: composition and quantity of income, position with regard to relations of production, social and cultural practices in both professional/productive or private/’non-productive’ contexts, habitus/class consciousness. It is important to take into account the polysemy of the notion of encadrement, the French meaning used here corresponding in English to management and supervision (the ‘cadre’ is the ‘executive’, that is, an individual with senior managerial responsibility) but also to the action/ activity of framing, as in ideological engineering and coordination. Obviously, this opens up a new area of enquiry, which is precisely where this research had led us: with few exceptions, the ‘sharing economy’, ‘commons’, and ‘platform cooperativism’ agitators appear to be spearheads of this encadrement class, spreading the word among other members of their class, consolidating deterioration of labour conditions for the working class, yet with some unavoidable ‘collateral damage’ within their own social group.

Chapter 5

Conclusions and Research Agenda for the Future Introduction In this book, we have first removed the noise that obfuscated the debate on the sharing economy by providing heuristically useful and empirically grounded conceptualisations. Second, we have reconstructed, deconstructed and unpacked rhetorical discourses and controversies. Third, we have mapped them against the available empirical evidence, including the most extensive review of secondary sources to date, as well as primary sources such as the wide analysis of platforms and the more in-depth fieldwork presented in Chapter 4. Fourth, we have considered legal disputes, the debate on regulation and policy and reviewed selectively some recent developments in this domain. We have done this by focusing on both the sharing economy in general, which from a regulatory and policy perspective mostly concerns consumers’ protection issues, and, more specifically, on digital labour markets and their implications on labour regulation and policy. In this conclusive chapter we will present both general considerations and others specifically related to consumer-oriented platforms or digital labour markets. First, we recall some of the main rhetorical themes and contrast them with conceptual and empirical analysis presented in the previous chapters. Second, we discuss again the regulatory and policy debate and highlight the open issues. Third, going back to the interpretative framework presented in the Introduction, we develop some brief reflections on the role of evidence in policy making and propose a few directions for the future research.

From Rhetoric to Conceptual Clarity and Evidence As the ‘sharing’ drama unfolded and started to be the source of conflicts, policy makers and regulators found themselves in front of new activities that blurred the boundaries between the personal and the commercial. They faced the challenge of ensuring the rule of law and fairness, protecting consumers, preserving labour rights and preventing the erosion of the tax base, while they were bombarded

Platform Economics: Rhetoric and Reality in the “Sharing Economy” Digital Activism and Society, 151–168 Copyright © 2019 by Emerald Publishing Limited All rights of reproduction in any form reserved doi:10.1108/978-1-78743-809-520181006

152    Platform Economics by discourses about the innovation potential of these platforms and how they should avoid stifling it. But they could hardly define clearly the boundaries of this new phenomenon, given that rhetorical framing and conceptual obfuscation are genetically attached to the emergence of these platforms and their self-defining practices. We have shown that commercial platforms self-defining themselves as part of sharing economy are, in fact, two-sided markets, and in some instances when they exert strict control on the trade, they resemble vertically integrated firms. They present almost no elements of sharing and simply enlisted the sharing rhetoric with its value-loaded connotation. Whereas bottom-up some participants may genuinely believe in the values of social exchanges and more community-like experiences, which lead, especially sociologists, to still put sharing platforms in relation to the moral economy, the platforms are only about making profit. Their business models (fee for transactions or fee for services) thrive only by increasing volume, and to this purpose consumers’ and workers’ rights are not always and necessarily protected. The various expressions used to refer to ‘sharing’ platforms have been and are used as ‘floating signifiers’ for all sorts of different activities in what can be called the rhetorical politics of platformisation. A closed selfreproducing loop has existed between conceptual ambiguity, rhetorical controversies and lack of sound measurements and empirical evidence. This loop, in turn, limits the possibilities of a rational debate of alternative policy options and contributes to the fragmented regulatory approaches, which currently address the ‘sharing economy’. We have contributed to breaking this loop by deconstructing and contextualising rhetorical discourses and by proposing typologies to guide both policy and the future research. The ‘sharing’ movement emerged as a form of social utopianism out of the broader narrative on the wisdom of the crowds and the creativity of the commons. After the development of ‘sharing’ platforms took a more ‘commercial turn’, disenchantment has fuelled growing criticism. Other more tangible interests (i.e. by disrupted incumbent industries) and concerns (i.e. by some policy makers, but especially by consumers’ advocated and trade unionists) exacerbated the conflictual debate that currently surrounds the ‘sharing economy’. These include, besides the political activation of the disrupted incumbents, urban tensions concerning the negative externalities caused by ride services and short-term accommodation rentals. The fact that ‘sharing’ platforms operate in a ‘grey area’ where they are neither legal nor illegal, but at times violate locallevel ordinances, also raises genuine or instrumental concerns about consumer protection and the rights of independent on demand workers. Combining the general analysis and that of digital labour markets, we have identified the following seven rhetorical themes: (1) Platforms helping to revive communities by strengthening social capital and increasing generalised trust. (2) Benefits trickled down, especially to the needy. (3) Promises of greener consumption (positive environmental effects) and wide net socio-economic welfare gains.

Conclusions and the Future Research Agenda    153 (4) Rhetoric of a flat world allowing digital labour migration with no boundaries and a world online meritocracy, which some digital labour markets derived from the work of economists and used in their public relations documents. (5) Extra money as a motivation for flexers to work in digital labour markets (students, retirees, stay-at-home parents etc.). (6) The alleged contribution of digital labour markets to bringing back to work the unemployed and under employed. (7) Discourse on the flexibility, autonomy and creativity that these platforms allegedly provide. (8) Illusionary and optimistic ideological production that technological advances and alternative digital labour organisation based on innovative models of production (commons/platform cooperativist) can ‘transform’ capitalism as model for the betterment of community and a fairer society. In addition to these rhetorical themes we found in media, blogs and in platforms’ investor relations documents, for what concerns digital labour markets, there is also the broad market efficiency hypothesis found in the economic literature that is very close to some of the rhetorical discourses. We refer to the hypotheses discussed in Chapter 3, according to which digital labour markets: (a) bring about a flat world online meritocracy; (b) increase labour market efficiency and (c) contribute to productivity growth. The empirical evidence found and analysed, despite limits that have been transparently acknowledged in previous chapters, enabled to shed light and deconstruct empirically some of these rhetorics and hypotheses. Although on some themes and hypotheses the available evidence is not conclusive to discard them, neither these can be used to confirm and push such narratives that are illegitimately pushed onto the regulatory and policy debate.

Evidence General Going beyond the polarised rhetoric and controversies, on the basis of the reviewed evidence, it can be stated that the ‘sharing economy’ overall is at best mixture of ‘passions’ and ‘interests’, although the latter prevail. Contribution to social capital, trust and community revival is possibly a feature of small, not-forprofit platforms. Utilitarian motivation to participate prevails when considering commercial platforms. There is no conclusive evidence supporting the claim that sharing platforms have a positive net welfare effect with distributional effects more pronounced for disadvantaged groups; having these positive and negative effects is an empirical question that cannot be answered by theory. On this aspect and on other socioeconomic impacts, there are only about a dozen or so methodologically robust studies, whose findings cannot be considered conclusive. Beside these studies, the rest of the evidence is simply anecdotal and often presented by stakeholders involved in the current controversies. For example, Uber and Airbnb have released dozens of reports about their positive impacts, but their reliability could

154    Platform Economics not be validated independently because the methodologies are not transparently illustrated and the data are not made accessible to all researchers. Equally inconclusive is the evidence on the promised positive environmental impacts of the ‘sharing economy’. It is extremely challenging and complex to demonstrate at aggregate level the net impacts in terms of environmental sustainability. First-order effects can reasonably be expected to be positive: staying in existing spaces would reduce the construction of new hotels and/or work spaces, while sharing tools or goods would reduce the production of new goods, both of which should reduce the ecological and carbon footprints. However, a measure of net impact at aggregate socio-economic level should also consider the secondorder effects. What happens with the extra-money that providers earned or users saved with the ‘sharing economy’? As seen, Airbnb has published ‘evidence’ that their hosts spend more than traditional tourists to show its impacts on city economies. This is self-defeating with respect to the claim of producing environmental benefits. The discourse on pin money and flexers being the main players is just a myth. Money is a key motivation, students and retirees are not the predominant groups. That these digital labour markets are increasing work force participation and bring back to work the unemployed is certainly not confirmed, and we would dare say it is refuted. In this respect, two qualitative studies (not included in the formal review, given their very exploratory nature and the use of very small samples) are worth mentioning (Dillahunt & Malone, 2015; Jen, Kaur, De Heus, & Dillahunt, 2014). They suggest that individuals from the most socially excluded social groups are not aware of these digital labour possibilities, and do not have the skills to participate in them. Possibly, there is the exception of female participation. The general hypothesis is that scheduled flexibility and the possibility of working remotely offered by digital work can help women, previously out of the labour markets, to ‘opt back in’ while managing other responsibilities, such as childcare (Dettling, 2016; Rossotto, Kuek, & Paradi-Guilford, 2012), or to overcome some of the cultural barriers that may exist in traditional workplaces (Raja, Imaizumi, Kelly, & Paradi-Guilford, 2013). The evidence is mixed as to whether female participation has increased in mobile labour markets (MLMs) for the provision of generic services. In TaskRabbit, women are overrepresented (Cullen & Farronato, 2015), but in other platforms (i.e. Gigwalk), men are overrepresented (Musthag & Ganesan, 2013; Teodoro et al., 2014). It is also debatable whether these digital markets do in fact provide the flexibility, autonomy and work–life balance that advocate and companies claim are the benefits of the gig economy. The impression given by the investigative journalistic reports is that on-demand work also involves dependent micro-earners and not only flexible and autonomous freelancers. The phenomenon of multiactivities further shed doubts on this hypothesis as does the surprising evidence of the engagement of full-time employees in what are second and third jobs. On the side of critical discourses, the data does not fully confirm that working in digital labour markets is an involuntary choice. Especially for those not positioned at the bottom of the society, it may be a deliberate choice. More flexibility and the possibility to work from home, on the other hand, may indicate that

Conclusions and the Future Research Agenda    155 this enables individuals to hold more than one job out of necessity. On the other hand, although not conclusive, the evidence on earnings, social protection and work conditions seem to support the claims that digital labour markets intermediate work that is even less guaranteed than in traditional non-standard work (NSW) and create precarity and inequalities. Most economic hypotheses about the increased efficiency produced by digital labour markets are not fully confirmed and, although more evidence is needed, available studies cast doubts about the broadly defined market efficiency hypothesis. First, it can be reasonably concluded that the world of online labour markets (OLMs) is not flat and is still far from being a globalised digital meritocracy. OLMs favour international labour flows (especially ‘North–South’) but are not as flat and meritocratic as expected. Various barriers that can be aptly summarised by the expression ‘the liability of foreignness’ (Lehdonvirta et al., 2014) limit the globalised trade of digital labour and the expected wage convergence; nonwesterners receive only a limited wage premium (compared with their domestic markets), but domestic contractors earn more in absolute terms, and are preferred for some tasks regardless of qualifications. This means that OLMs are less beneficial than expected for developing countries, but also that they exert less pressure in driving down wages in more developed countries. Second, the evidence on matching frictions and hiring biases (or inefficiencies) is very extensive and solid, including several experimental studies. It shows that these markets are still not as efficient as expected and that Autor (2001, 2008) had a strong point when he voiced scepticism about ‘wired labour’. There is a fairly large number of experimental and quasi-experimental studies showing that digital labour markets are still ridden with matching frictions and hiring inefficiencies. Little correlation is found between skills and earning levels, whereas reputational ratings and references seem to be the main explanation for the jobs and money contractors manage to secure. It seems that referral information is used by employers more than all other observable characteristics on which information is fully available in OLMs. There is a risk that such frictions and hiring inefficiencies may exacerbate wage inequalities by further skewing work in favour of the most skilled and precluding entry by inexperienced workers. In this respect, it is important to note the relevance of also applying a behavioural perspective to the documented cases of gender- and ethnic-based discrimination resulting from, in most cases, involuntary bad decisions based on stereotypes as heuristics, leading to judgements and decision-making affected by confirmation biases. Third, evidence support only superstar effects, which suggest that polarisation rather than equalisation is produced. Finally, so far there is no evidence of aggregate net effects. Although not related to rhetorical themes, we analysed digital labour markets in a broader perspective and showed how they are a further form that NSW is taking. Their potential growth in the future depends not only on technology and the future of work but also on the institutional trend towards de-standardisation of labour and job polarisation with the hollowing out of standard middle-level jobs. Routine tasks requiring middle skills are among the most traded digitally and are performed by individuals with the same profile as those

156    Platform Economics being laid off or not hired from firms under regular work forms, which means they are outsourced not just because of technology (as assumed in the ‘routine biased technical change’ hypothesis) but also because it is institutionally possible and economically more convenient. So, while currently limited, if digital NSWs continue to grow, it is an empirically consequential hypothesis that they could potentially encroach traditional and long-term forms of employment (Einav et al., 2015, p. 20). With respect to the issue of NSW, we have shed light on at least four important aspects: employment status, earnings and social protection, voluntary or involuntary choices and working conditions. On the employment status and history of those working in digital labour markets, much more research is urgently needed and the evidence reviewed is fragmentary and mostly indirect. Berg (2016), for instance, using only two small samples reports that 33% of those working for micro-tasking OLMs, such as MTurk and Crowdflower, were unemployed, whereas the remaining two-thirds use these markets to complement part-time or full-time employment. Other surveys have shown that providers of labour-intensive services are either self-employed or are full-time employees with open-ended contracts. The engagements of full-time employees are second and third jobs, together with the increasing phenomenon of multiactivities challenge the idea that these markets are used to balance life and work and that they provide opportunities for underemployed. Evidence is fairly uniform in attesting that earnings in such markets range from very low to modest with only a small minority of workers making above middle-level incomes and that workers have no form of social protection, are in a position of unfavourable information and power asymmetry, and that their privacy is not protected. Gender- and ethnicity-based discrimination (voluntarily or involuntarily produced by matching frictions, hiring inefficiencies and cognitive biases) is not uncommon and workers have no way to protect themselves from it. Several studies have documented the increasing diffusion of strictly automated control through algorithms, which has been further corroborated by the inquiries of US judges in the legal cases involving Uber and Lyft. This seemingly support Cherry’s (2006) claim that digital labour markets bring us back to Taylor (automated control), Smith (division of labour in pieces), and pre-industrial levels of work precarisation (lack of social protection). The question whether these forms of work are voluntary or involuntary is a thornier one, and the evidence available does not enable any conclusive statement. At the opposite extreme, some claim that from the fast pace of growth in the number of contractors and their profile, it can be derived that working in OLMs is a free choice (Agrawal et al., 2013a), whereas others using qualitative in-depth ethnography conclude that it is not a totally uncoerced choice (Irani, 2015, p. 227). In a few surveys, respondent complain about low pay and not sufficient and steady flow of work. Obviously, there is a wide array of situations ranging from freely choosing independent freelancing to doing gigs for lack of alternative opportunities.

Conclusions and the Future Research Agenda    157

Policy and Regulation: Debate and Open Issues General Issues In the midst of controversies and legal disputes, the various regulatory essays reviewed are to some extent polarised between those radically against any intervention (Allen & Berg, 2014; Cohen & Sundararajan, 2015; Koopman et al., 2014, 2015; Sundararajan, 2014; Thierer et al., 2015) and those that are in favour of some forms of regulation (Edelman & Geradin, Forthcoming; Gobble, 2015; Malhotra & Van Alstyne, 2014; McLean, 2015; Ranchordas, 2015; Rauch & Schleicher, 2015; Sunil & Noah, 2015; Zrenner, 2015). There are also some more specialised legal approaches (Barry & Caron, 2014; Cohen & Zehngebot, 2014; Daus & Russo, 2015; Miller, 2014, 2015) proposing very strict interventions in, for example, transportation services (Daus & Russo, 2015). The libertarian solution uses the weaponry of textbook economics about the failures of regulation and the self-regulatory nature of markets (Allen & Berg, 2014; Koopman et al., 2014; Thierer et al., 2015). A key and objectively valid point is the risk of regulatory capture of regulators by incumbent industries. This dynamics can promote socially unproductive but costly rent-seeking behaviour by firms seeking to maintain their market stronghold through lobbying, donations and other means. From a libertarian standpoint, excessive legislation and regulation could absorb and neutralise the benefits to consumers and the efficiency gains allegedly produced by technological innovation. According to this perspective, the ‘sharing economy’ has allegedly overcome market imperfections without recourse to traditional forms of regulation. The Internet and the rapid growth of the sharing economy alleviate the need for much of this top-down regulation, and these recent innovations probably do a much better job of serving consumer needs. It is argued that the ‘sharing economy’s’ reputational feedback mechanisms solve the information asymmetry, commonly called the ‘lemons problem’ (Thierer et al., 2015).1 From this perspective, a new approach to bottom-up self-regulation is needed where various forms of licencing should be reduced to allow private certification schemes and reputation mechanisms to evolve; regulations making it difficult for start-ups to compete for labour (contractors should not be turned into employees) should be avoided; and regulation should remain general and not industry-specific. More nuanced and less radical approaches call for innovative and smart forms of regulation, which attempt a compromise to ensure consumer protection and safety without stifling innovation (Barry & Caron, 2014; Miller, 2014, 2015; Ranchordas, 2015; Rauch & Schleicher, 2015; Sunil & Noah, 2015). By and large, the smarter regulations envisage a number of possible 1

In a famous paper, Akerlof (1970) describes how information asymmetries prevent certain mutually beneficial exchanges from taking place. Considering the used car market, he explains that used car buyers know that ‘lemons’ (bad cars) exist but are unable to distinguish them from higher-quality cars, and they are therefore less willing to pay. The buyers’ uncertainty, in turn, discourages sellers of higher-quality cars from offering their cars for sale, making both buyers and sellers worse off.

158    Platform Economics solutions: use of information-based regulation (metrics and performance); development of a general but differentiated regime for the ‘sharing economy’; co-opting of the ‘sharing economy’ organisations into the city governance structure, as was done in the past with industries that performed a quasipublic service; not applying traditional regulation to the ‘sharing economy’ but rather, if necessary for the sake of fair competition, gradually deregulating incumbent industries. Sunil and Noah (2015), for instance, recommend that governments should establish a strategic operating framework, realign political and cultural incentives, and modernise their structures to be ready to manage smart regulation regimes. Importantly, they also stress that the sharing economy companies will have to make their data fully open if a regime like this is to emerge. Rauch and Schleicher (2015) interestingly observe and then challenge the current sharing wars, for they rely on an unstated assumption: If the sharing firms win these fights, their future will be largely free from government regulation. Local governments will either shut sharing down or they will leave it alone. However, they envisage that as the ‘sharing economy’ firms move from being start-ups to being important and permanent players in key urban industries (transportation, hospitality and dining), local and state governments are likely to adopt the mixed regulatory strategies they apply to the types of firms with which sharing firms share important traits, for example, property developers and incumbent taxi operators. In this spirit, a very technical but interesting solution has been suggested by Miller (2014) for short-term rental market in general and Airbnb in particular. He proposes a ‘transferable sharing right’ (TSR) mechanism, which is modelled on existing transferable development rights regime. This TSR regime would provide cities with a means of regulating short-term rentals while charging a fee equal to the resulting externalities and lost city revenues. Furthermore, TSRs could be used to reinvest in neighbourhoods where shortterm rentals occur, or to drive economic development to neighbourhoods where cities seek to encourage tourism. Cannon and Chung (2015) argue in favour of a co-regulation approach, as certain areas of the ‘sharing economy’ are suited to regulatory intervention and others to self-regulation. They warn, for instance, that when both suppliers and consumers depend on one another for reviews, there is a risk of retaliation, which can lead users to soften negative reviews and make ratings less negative and thus less reliable. They also underscore the need for introducing a minimum insurance requirement, as imposed by California, for instance, on ride-sharing companies. A couple of very balanced appraisals mix economic analysis and regulatory considerations. They argue that old-fashioned and ineffective regulations should not stifle innovation and undermine efficiency gains and consumer welfare gain and they recognise that platforms cannot continue from above the law as they currently do (Edelman & Geradin, Forthcoming; Einav et al., 2015). Edelman and Geradin (Forthcoming, pp. 9–13) consider imposing licencing schemes on ineffective platforms and a source of regulatory capture by incumbents. However, they urge platforms to be ready to accept requirements which genuinely protect both customers and non-customers. With respect to the latter,

Conclusions and the Future Research Agenda    159 they stress the need for intervention when platforms’ negative externalities affect non-customers who lack any contractual relationships with platforms or service providers. Einav et al. (2015) recognise that ratings can be biased and inflated and that it is possible that platforms present the results of search in a way that is more convenient to them than to the users. On the other hand, they also point out that imposing licencing and certification on the platform may protect incumbents without really protecting consumers. Although these requirements can be seen as remedies to market failures, their implementation takes the form of lengthy processes, after which little monitoring is performed. In this respect, they seem to favour small interventions, which allow traditional industries and new platforms to compete on an equal footing. With respect to the use of data by the platforms, they observe that several questions emerge, such as the following: Can consumers limit the use by platforms of data? Can platforms share/sell ratings and purchase history? What about potential gender and race discrimination in ratings leading to certain groups getting fewer opportunities? Regardless of the regulatory debate described and despite the availability of various proposals, in practice, regulatory regimes are lagging behind. As a result, many ‘sharing’ activities occur in the ‘grey area’ and leave open a number of issues: (1) Taxation. Substantive laws exist for tax-sharing activities. However, enforcement may present challenges because (a) some platforms opportunistically pick the more favourable regulatory regime; (b) micro-providers raise unique compliance concerns. Airbnb is currently engaged with legislators in drafting or adjusting existing legislation. In addition, its website informs hosts about local laws and their landlord’s rental policies, and requires hosts to comply with them, both of which may prohibit short-term rentals (Miller, 2015; Zrenner, 2015). Furthermore, Airbnb has also started to collect taxes in some US cities and Amsterdam. (2) Negative externalities, liability and insurance. Negative externalities for ride service platforms are derived from unsafe and uninsured or under-insured drivers/cars. Short-term accommodation rentals produce negative externalities on neighbourhoods (increased traffic, parking place occupied, noise, tenants disturbing neighbours, etc.) and by removing properties from long-term rental markets. Liability and insurance, however, are not only a matter of negative externality but may also concern the two sides of a ‘sharing’ transaction. The issue is again to determine as to who is liable if something goes wrong and to guarantee that ‘sharing’ activities are insured. It is reasonable to expect that some intervention may be needed to define liability, ensure safety, and close the insurance gap. Under specific circumstances, the negative externalities of short-term rentals should also be addressed. (3) Information asymmetries and cognitive biases. Various information asymmetries, exacerbated by the typical cognitive biases documented in behavioural economic literature, cast doubt on the extent to which self-regulation fully protects consumers. This entails various specific issues such as the

160    Platform Economics reliability of reputational ratings, safety standards, fraud, dispute resolution and redress. The chances are that consumers will make poor decisions when faced with an overwhelming range of choices, poor regulation and unclear avenues for recourse in the case of a dispute, not to mention the fact that they may fail to fully appreciate risks and safety requirements. Under these circumstances, regulation and/or nudges could help increase consumer protection. (4) Licencing and certification schemes. Licencing and certification schemes tend to be ineffective and may unduly favour incumbents. However, instances of serious incidents with both Uber and Airbnb have caused critics to call for imposing licencing and certification on large commercial platforms. Platforms try to boost confidence with identification (ID) checks and vetting processes, but there are doubts on the transparency and rigorousness of these inspections. (5) Data and privacy. There are concerns about the amount of data that ‘sharing’ platforms are collecting about consumers, given the sensitive nature of some of these data and how these are being used. (6) Competition law potential implications. From the evidence reviewed on the characteristics of the largest platforms and how they function, it seems that market dominance is out of reach for the most of these platforms due to heterogeneity and matching frictions. However, it is not so unlikely for Uber. On the other hand, improvements in the matching algorithms, together with pricing strategies and use of personal data without any regulatory checks, may change the situation and make market dominance more likely for a few other platforms. In the review by Lougher and Kalmanowicz (2016), it is not ruled out that ‘sharing’ intermediation markets can become concentrated and possibly dominated by a single market player. The activities of powerful platforms, for which data use is key, are likely to be scrutinised in merger control proceedings, and in the long term potentially also in the area of market abuse. They cite statements by representatives of the French and Germany competition authorities to substantiate the claim that market power for such platforms comes from the capacity to collect a large amount of personal data and use it commercially (Lougher and Kalmanowicz, 2016, pp. 96–97). Finally, after noting that the regulation of ‘sharing’ platforms is hotly debated, they report that whereas some member States have called for specific regulatory framework, however the European Competition Commissioner Margrethe Vestager has made clear in several statements that such platforms are too diverse to be monitored through a single regulatory framework and that it is preferable to apply existing antitrust rules case-by-case (p. 102). From the evidence reviewed on the characteristics and functioning of the largest platforms, it seems that market dominance is out of reach for most of them due to heterogeneity and matching frictions, but is not so unlikely for Uber. On the other hand, improvements in the matching algorithms, together with pricing strategies and use of personal data without any regulatory checks, may change the situation and make market dominance more likely also for a few other platforms.

Conclusions and the Future Research Agenda    161 Labour-specific Issues There are four broad questions to which policy makers would certainly like to have firm answers backed by robust evidence: (a) What are the possible implications of these new digital labour markets on employment and wages? Do they create new jobs or simply crowd out existing ones? Are they a source of income integration for the underemployed, or are they rather contributing to downward pressure on wages? (b) Do they justify a regulatory intervention? If yes, in what areas (i.e. taxation, liability, insurance and social protection)? (c) What would be the costs of curbing innovation and loosing on improved labour market efficiency because of regulatory intervention? (d) Are there risks in Europe that fragmentation will emerge as a result of national or local interventions, or in cases where the issue of classification (self-employed vs workers) will be decided by the courts in the absence of regulation? Obviously, the evidence collated in this book can provide only tentative and partial answers, and it is only by filling in the gaps already outlined and further discussed in the concluding paragraph that firmer answers will be possible. First of all, more surveys and data are needed to better assess the dimensional relevance of this phenomenon and its possible developmental paths because this will determine the extent to which the above questions are really policyrelevant. Certainly, if growth continues at a fast pace of the last five years, these new markets could encroach on traditional and long-term forms of employment. Currently, however, the size of these new markets is limited. Economic theory suggests that there will be both further growth (reduction of search, coordination and transaction costs) and countervailing effects (frictions, cost of quality control, etc.). With regard to question (a), the evidence is inconclusive and lends itself to different interpretations in terms of potential impacts on labour market dualism, employment polarisation and income inequality. In other words, there is still ambiguity on the direction of the effects and, especially, about how these effects will be distributed both within and between countries. It has been suggested that digital labour markets can be together with more general form of NSW, the other side of the job polarisation story told by the ‘routine biased technical change’ literature. This said, the evidence reviewed is absolutely insufficient to conclude that this is the case, and discern whether these markets increase the dualism of the labour markets or generate a polarisation between these new forms of flexible work and regular employment (full-time, but also standard part-time and fixedterm employment). Lack of evidence also prevents us from assessing the extent to which firms outsource non-core or core tasks to these markets, which is a key to understand the main direction of the future development. On the other hand, question (b) is relatively easier to answer. The evidence shows that the amount of money workers can make on these platforms varies widely from very little to

162    Platform Economics just above the minimum wage, and that work–life balance and working conditions are far from ideal (no social protection, asymmetries, surveillance, lack of privacy protection, etc.). While caution must be exerted, especially on the classification issue (need to avoid treating truly independent freelancers as dependent self-employed), there is enough evidence and several reasonable proposals for regulatory intervention. In this respect, it is fairly clear that platform liability should be better defined both in general and with specific regard to third party damages and to accidents workers may have as they perform tasks (this mostly concerns MLMs). The answer to question (c) on what society stands to lose, if regulation curbs labour market innovation, finds the same limitation as question (a), due to lack of conclusive evidence on key effects. Answering this question would require a cost–benefit analysis weighing in positive and negative effects, but empirically the evidence is not conclusive for either. Positive effects identified as ex ante (production efficiency, aggregate welfare effects from more efficient labour markets, productivity and indirect employment gains, increased participation of inactive and unemployed) are not empirically confirmed, yet should not be discarded. Finally, on question (d), it should be recalled that this book did not perform a review of regulatory developments in EU28 and that this undertaking could be an important complement to the evidence presented here. On the other hand, it is fairly clear that if in Europe things develop as they have gone in the United States and arising issues are defined by court decisions, this will cause fragmentation. It is, thus, urgent that some EU-level guidance be provided so that member States can introduce some form of regulation to reduce any potential fragmentation. More generally, policy makers should acknowledge that employment in the twentyfirst century is no longer a binary phenomenon (1 = employed, 0 = unemployed) and set a target to minimise involuntary employment and under-employment (Atkinson, 2015, Chapter 5) while letting open the opportunity for flexible and small pieces of jobs to be performed by those freely choosing to do so. They should put in place fair conditions for voluntary atypical work so that it does not increase inequality. In the same vein, it has been argued that a regulatory approach should not be 1 = employees and 0 = contractors, but ways should be found to increase protection without suddenly increasing the costs for digital platforms. The second wave of digital transformation could be beneficial as long as researchers, policy makers, trade unions and industry find innovative institutional ways of helping to exploit the opportunities without neglecting the social challenges (Brynjolfsson & McAfee, 2014; Brynjolfsson et al., 2015). The ‘Open Letter on the Digital Economy’ (Brynjolfsson et al., 2015) calls companies to ‘develop new organisational models and approaches that not only enhance productivity and generate wealth but also create broad-based opportunity’. The European Economic and Social Committee (EESC, 2014) in a draft opinion stresses that new policy measures and new agreements on the organisation of work are needed to avoid a situation whereby digitalisation further increases inequality, reduces job quality, and worsens working conditions. As shown, industry leaders and foundations have followed up with the proposal to make social benefit portable across gigs; although not yet sufficient, this would already represent a step ahead.

Conclusions and the Future Research Agenda    163 Our View on Labour Issues While platforms might help in providing more and better job opportunities for some workers (hence improving fairness), they also fuel the growth of non-standard forms of employment that, in many cases and if unregulated, imply worse working arrangements and conditions compared with those experienced by workers with standard employment contracts (with a negative effect on fairness). The coexistence of positive and negative effects is confirmed when one looks at the relationship between the development of labour platforms and inequality of wages: while in some aspects – mostly related to the efficiency-enhancing aspects of labour platforms – these might lead to a reduction in wage inequality, for others (such as the superstar effect), these are likely to increase it, so that the response can only be provided on an empirical level (while in the case of fairness, the policy response can be offered just with an ex ante analysis of the differential treatments under standard and non-standard working arrangements). Notice also that fairness and equality are linked tightly as ex ante input and ex post outcomes and should be tackled through both policy and regulation, since the inequality outcomes of today will become the unfair starting conditions of tomorrow (Atkinson, 2015, Chapter 1). People deserve to keep a reasonable portion of what they earn through increased hours or taking increased responsibility for a second job. Principles of social justice require that individuals have access to primary goods such as rights, powers and income (Rawls, 1971), not forgetting that people also have very different capacities of converting primary goods into a good living (Sen, 2009). Ex ante opportunities are better linked to ex post outcomes by expanding the relevant dimensions in terms of capacity and functioning (Sen, 1999). Labour markets are social institutions, since there is something special about labour as a commodity, including the fact that participants on both sides hold well-developed notions and norms about what is fair and what is not (Solow, 1990). Social norms and the notion of fairness actually remove indeterminacy since individual incentives are not sufficient to reach a unique equilibrium in the market (MacLeod & Malcomson, 1998). Social norms and values can be consistent with agent rationality and shape economic behaviour because of, for instance, their implications for the reputation and public legitimacy of workers and employers (Solow, 1990). In view of the above considerations and the empirical evidence reviewed, it is not unreasonable to expect policy makers, regulators and social partners in Europe to meet and define a consensual roadmap towards the establishment of a fair and dignified support infrastructure (FDSI) for on-demand workers that does not jeopardise innovation. As a relative late comer to the ongoing disputes and conflicts erupting in the United States, Europe could learn from this experience and avoid issues ending up in court and becoming radicalised. European stakeholders should also learn from the past and acknowledge that the dual policy approach of liberalising new forms of employment while retaining standard full-time employment as a benchmark was not fully effective and has to some extent exacerbated labour market dualism. An FDSI should ease convergence and transition between digitally mediated on-demand work and other forms of employment, limiting as far as possible the exemptions and exceptions that have undermined the effectiveness of previous directives and social partner agreements.

164    Platform Economics This support infrastructure should include the following pillars: ⦁⦁ A minimum wage should be defined together with limits to the maximum num-

⦁⦁ ⦁⦁ ⦁⦁ ⦁⦁

ber of hours worked per day (acceptance rates should not be used by platforms to deactivate or terminate workers’ contracts). Some minimal forms of social protection and health insurance should be introduced. Liability insurance for damage to third parties should be considered, and some forms of health-safety measures. The kind and frequency of technological forms of control and the use of workers’ data should be regulated to ensure the protection of privacy. It should be ensured that in maximising volumes for the platforms algorithmically automated matching and reputational ratings do not produce discrimination with respect to gender, ethnicity, race and age.

More generally, the FDSI should facilitate individuals’ access to the standard forms of employment at certain points during their lives and then to more flexible forms of work at other points. Individuals should not be penalised for these transitions by the loss of seniority rights and occupational benefits. So far, in Europe, existing settings facilitate the transition from standard to non-standard forms of employment, but not vice versa. Although it is difficult to identify and measure involuntary atypical digitally mediated work, policies should minimise this type of work while, at the same time, allowing voluntary adhesion to flourish under conditions of fairness and dignity.

Lobbying as Rhetorical Framing: The Role of Evidence and a Research Agenda for the Future This book has amply documented the fact that the ‘sharing economy’ is currently characterised by conflicting rhetoric and controversies between disputed values and interests. Factual evidence is currently limited, and yet evidence has been instrumentally used in what in the Introduction we called lobbying as rhetorical framing. The instrumental rhetorical discourses, when empirical evidence is still lacking or inconclusive, have contributed to the undersupply of policy and regulatory responses and today to policy and regulatory decisions taken ‘in the dark’ often under the influence of some interest groups. But what emerges clearly is that evidence has become the main currency of lobbying, and commercial platforms have used it effectively to produce a negative policy bubble. Some academic economists have actually helped commercial platforms directly by using data made available only to them to present partial analysis about the social benefits produced. Others, in the face of insufficient evidence and epistemic uncertainty, have not hesitated to extol the virtues of these platforms. In the case of ‘sharing platform’, more solid evidence is needed for sound regulation and policy as common also in other policy domains. Epistemic uncertainty is a condition of science and cannot be removed, and research efforts are always needed to construct a more robust evidence base for policy. These efforts cannot

Conclusions and the Future Research Agenda    165 promise to solve all the conflicts and controversies following a pure ‘technocratic’ model. They could, however, more ‘realistically’ and ‘humbly’ reduce the current ‘value-loadedness’ that characterises not only the public debate but also many of the more academic and supposedly scientific contributions. A number of essential areas where gaps are evident and research is needed are presented in this section. Before doing this, however, it is worth spending a few more words on the relation between science and policy briefly introduced in Chapter 1. In many cases, the use of scientific advice from a ‘pure scientist’ (disinterested in policy-making process and simply providing neutral information) or a ‘science arbiter’ (adjudicating claims through scientific research as part of panels or advisory boards) would amount to ‘stealth issue advocacy’. It is worse when the legitimacy of a scientist is used to reduce the scope of choices available for policy, as seen in some of the reports commissioned by Airbnb and Uber. This actually turns scientific work into issues’ advocacy (Fig. 9). The only possible way to progress when problems are intractable is to provide scientific expertise in the spirit of what Pielke (2007) calls the ‘Honest Broker of Policy Alternative’. This approach is based on clarifying and possibly widening the choices available to decision makers. Indirectly or directly, this approach facilitates the integration of scientific advice with the perspective and input of all the stakeholders involved. In particular, this is important when evidence is used as the main currency of lobbyists and when some experts pretending to be pure scientists or science arbiters in reality are issue advocates. In view of these considerations and of all the contents of this book, below we point to some elements for a European research agenda specific to digital labour markets that would fill some of the existing gaps. As the first step towards a research agenda in support of policy-making, the following key research questions need to be answered: (a) What is the direct employment effect of labour market platforms, that is, how many workers are actually involved in the functioning of labour market platforms?

Fig. 9:  Determining the Role of Science in Policy and Politics. Source: Adapted from Pielke (2007, p. 19).

166    Platform Economics (b) Who are the service providers (i.e. on the supply side)? What are their demographics, education, location and motivation, and how do they judge the experience? This will also help us to answer questions about the substitution of regular labour contracts with labour supplied through platforms from the perspective of those supplying the labour. It will also provide a better understanding of their opinions on the conditions of employment in the sharing economy (hours of work, wages, quality of work, etc.). (c) From the user’s perspective (i.e. the demand side), what kinds of tasks are outsourced? Are outsourced tasks mostly related to non-core activities and/ or standardised activities (where the ‘buy’ option dominates over the ‘make’ one)? Are these tasks mostly cognitive, manual or interactive? Are they routine or non-routine tasks? What are the main drivers of firms to use labour market platforms? A specifically delimited and short- and medium-term research project could include the following steps: (1) Select one of the types identified in the matrix. (2) Conduct a more focussed review of literature on this type, striving as far as possible to capture non-English language papers and grey literature focussing on European settings. (3) Perform a wide-ranging web-based review of digital labour markets belonging to chosen type and active in Europe. (4) Based on the previous steps, make a preliminary identification of those that seem to have achieved some scale and, through short telephone interviews, obtain more data to validate the selection. In spite of the fact that these markets could be international or global, data on the labour platforms should be collected locally, that is, based on their prevalence in a given country (hence, we would have about 20 country-specific markets, covering five major European countries such as the UK, Germany, France, Italy and Spain). (5) Depending on time and budget, select 10 of these digital labour markets (two digital labour markets per country) for more in-depth analysis based on:

(a) in-depth, thick qualitative case studies in the field (interviewing representatives of the digital labour markets, and a few contractors and employers); (b) online surveys of contractors and employers to be administered through digital labour markets; (c) if possible, econometric analysis of primary data obtained by such markets (replicating some of the studies reviewed) and (d) design and realise field experiments (replicating some of the studies reviewed).

(6) In parallel to the above steps, launch a survey in the five major EU countries based on nationally representative samples to measure the ‘prevalence’ of employment in digital labour markets and to obtain data on workers’

Conclusions and the Future Research Agenda    167 socio-demographic profiles, employment status and histories, earnings, motivations, how they consider the choice between gig and other forms of work, and how they assess the experience of working for digital labour markets. Robust empirical studies are emerging, meaning that data are potentially available. The evidence reviewed reveals that there are ‘big data’ research opportunities, either in accessing data provided by the platforms or web scraping some parts of it. Economists have started to use big data (Einav & Levin, 2014; Taylor et al., 2014), and, more recently, a research agenda for ‘big data labour economics’ has been proposed (Horton & Tambe, 2015). Big data can be used to study, for instance, the geographic composition of the pool of contractors and further analyse the effects of language, culture, genders and other possible barriers on matching. There is, however, a possible drawback to the ‘big data opportunity’ represented by the rise of what have been called ‘embedded researchers’ (Ruths & Pfeffer, 2014). These are researchers with special relationships with platforms and access to their data, which create a divide in the research community and a little possibility of validating/replicating their results. This problem, which seriously affects research on the ‘sharing economy’ should be urgently remedied. As Cohen et al. (2016, p. 22) argue: this paper also points to a second path forward: one in which better data are the key to deeper insights. Massive changes that are taking place in the economy in terms of the availability of transaction level data, the increased use of sophisticated pricing tools by firms, and the growing openness of firms to randomized experiments. All of these forces point towards a future world in which data richness transforms our understanding of firms and consumers. Big data alone, however, would not fill all evidence gaps, and need to be integrated with more traditional survey data as well as with in-depth qualitative studies. One evident gap, for instance, concerns the measurement and quantification of the phenomenon. In this respect, more surveys based on nationally representative samples are needed, and should be complemented with a collation of descriptive data on more platforms than those for which data have been found and reported in this book. There are evidence gaps in most of the topics discussed, and there is still much to be done to sharpen our understanding of these new labour markets and their implications. There is a huge lack of evidence on European contexts compared with the United States, which is taken into account in the following considerations. First, the kind of experimental and observational analysis based on platforms data that have been reviewed should be replicated across different types of digital labour markets and across different locations, especially for what concerns localised MLMs. Studies on OLMs should be conducted with a more granular and extensive focus on the country of origin of both employers and contractors, and with differentiation by categories of tasks traded and by firms’ sectors and size.

168    Platform Economics The second gap regard to the ‘workers’ socio-demographic profiles, employment status and histories, earnings, motivations, how they consider the choice between gig and other forms of work and how they assess the experience of working for digital labour markets. This kind of information can be gathered triangulating traditional surveys, surveys on specific digital labour markets and qualitative case studies on the latter. Third, the evidence on distributional effects on employment and income is emerging but is still limited and inconclusive; it is of clear importance and policy relevance to ascertain whether ‘super star’ or ‘long tail’ effects are prevalent; studies on such effects should be replicated and expanded with a triangulation between administrative data from digital labour markets and surveys/in-depth interviews with their contractors. Information is needed on the country of residence and socio-demographic profile of contractors where super star effects emerge, as well as which type of firms (size and sectors) are more active in sourcing tasks to platforms; whether the outsourced tasks are part of core business, or are rather non-core tasks. The issue of control, algorithm management and working conditions need to be further documented through qualitative in-depth studies complemented with other studies that look more specifically at regulations and labour laws for these new forms of employment. More granular overviews are needed on: (a) how different digital labour markets deal with liability and insurance issues; (b) differences in European countries in the criteria applied to distinguish self-employed and workers and (c) whether new legal approaches are emerging on the latter issue with specific regard to sharing economy platforms. The matching process with its frictions is well documented, but studies should be replicated for more platforms and different locations. Finally, aggregate welfare effects are certainly important, but are probably the object of a more long-term agenda.

References Abdul-Rahman, A., & Hailes, S. (2000). Supporting trust in virtual communities. In HICSS 2000 proceedings of the 33rd Hawaii international conference on system sciences (Vol. 6, p. 132). Washington, DC: IEEE Computer Society. Adams, Z., & Deakin, S. (2014). Re-regulating zero hours contracts. Liverpool: Institute of Employment Rights. Adelman, J. (2013). Worldly philosopher. The odyssey of Albert O. Hirschman. Princeton, NJ: Princeton University Press. Adorno, T. (2006). History and freedom: Lectures 1964–1965. (R. Tiedemann, Ed. & R. Livingston, Trans.). Cambridge, UK: Polity Press. Agrawal, A., Horton, J., Lacetera, N., & Lyons, E. (2013a). Digitization and the contract labor market: A research agenda. NBER Working Paper Series 19525. doi:10.3386/w19525 Agrawal, A., Lacetera, N., & Lyons, E. (2013b). Does information help or hinder job applicants from less developed countries in online markets? NBER Working Paper Series, 18720. doi:10.3386/w18720 Agyeman, J., McLaren, D., & Schaefer-Borrego, A. (2013). Sharing cities. Washington, DC: Friends of the Earth. Retrieved from http://www.pando.sc/sites/default/files/ documents/agyeman_sharing_cities.pdf. Accessed on May 15, 2015. Airbnb. (2012). San Francisco. Retrieved from http://blog.airbnb.com/economic-impactairbnb/-san-francisco. Accessed on June 8, 2015. Airbnb. (2013a). Amsterdam. Retrieved from http://blog.airbnb.com/economic-impactairbnb/-amsterdam. Accessed on June 8, 2015. Airbnb. (2013b). Berlin. Retrieved from http://blog.airbnb.com/economic-impact-airbnb/berlin. Accessed on June 8, 2015. Airbnb. (2013c). New York. Retrieved from http://blog.airbnb.com/economic-impactairbnb/-new-york. Accessed on June 8, 2015. Airbnb. (2013d). Paris. Retrieved from http://blog.airbnb.com/economic-impact-airbnb/paris. Accessed on June 8, 2015. Airbnb. (2014a). Barcelona. Retrieved from http://blog.airbnb.com/economic-impactairbnb/-barcelona. Accessed on June 8, 2015. Airbnb. (2014b). London and Edimburgh. Retrieved from http://blog.airbnb.com/economicimpact-airbnb/-london-edinburgh. Accessed on June 8, 2015. Airbnb. (2015a). Airbnb’s positive impact in Athens. Retrieved from http://blog.airbnb.com/ airbnbs-positive-impact-in-athens/. Accessed on June 8, 2015. Airbnb. (2015b). Economic impact. Retrieved from https://www.airbnb.com/economicimpact. Accessed on June 8, 2015. Akerlof, G. (1970). The Market for ‘Lemons’: Quality Uncertainty and the Market Mechanism. Quarterly Journal of Economics, 84(3), 488–500. Alba, D. (2015, July 2). Shyp makes couriers employees before it’s too big to change. Wired. Retrieved from http://www.wired.com/2015/07/shyp-makes-couriers-employees-bigchange/. Accessed on January 21, 2016. Albinsson, P. A., & Yasanthi Perera, B. (2012). Alternative marketplaces in the 21st century: Building community through sharing events. Journal of Consumer Behaviour, 11(4), 303–315. doi:10.1002/cb.1389 Allen, D., & Berg, C. (2014). The sharing economy: How over-regulation could destroy an economic revolution. Melbourne, Australia: Institute of Public Affairs. Retrieved from

170   References https://ipa.org.au/portal/uploads/Sharing_Economy_December_2014.pdf. Accessed on May 21, 2015. Allmendinger, J., Hipp, L., & Stuth, S. (2013). Atypical employment in Europe 1996–2011. Discussion Paper No. 1169 P 2013-003. Berlin, Germany: WZB. Aloisi, A. (2015). Commoditized workers the rising of on-demand work, a case study research on a set of online platforms and apps. Retrieved from http://ssrn.com/abstract=2637485. Accessed on December 2, 2015. Anderson, C. (2006). The long tail: Why the future of business is selling less of more. New York, NY: Hyperion. Andrejevic, M. (2009). Exploiting Youtube: Contradictions of user-generated labor. In P. Snickars & P. Vonderau (Eds.), The youtube reader (pp. 406–424). Stockholm: KB. Andreoni, J., & Rao, J. (2011). The power of asking: How communication affects selfishness, empathy, and altruism. Journal of Public Economics, 95(7–8), 513–520. Aneesh, A. (2009). Global labor: Algocratic modes of organization. Sociological Theory, 27(4), 347–370. Angrist, J., Caldwell, S., & Hall, J. (2017). Uber vs. taxi: A driver’s eye view. NBER Working Paper Series No. 23891 (p. 19525). Retrieved from http://www.nber.org/papers/w23891. Accessed on May 2018. Armstrong, M. (2006). Competition in two-sided markets. Rand Journal of Economics, 37(3), 668–691. doi:10.1111/j.1756-2171.2006.tb00037.x Arntz, M., Gregory, T., & Zierahn, U. (2016). The risk of automation for jobs in OECD countries: A comparative analysis. OECD Social, Employment and Migration Working Papers, No. 189. Paris, France: OECD. Arsel, Z., & Dobsha, S. (2011). Hybrid pro-social exchange systems: The case of freecycle. Advances in Consumer Research, 39, 66–67. Atkinson, T. (2015). Inequality: What can be done? Cambridge, MA: Harvard University Press. Autor, D. (2001). Wiring the labor market. Journal of Economic Perspectives, 15(1), 25–40. Autor, D. (2008). The economics of labor market intermediation: An analytic framework. Cambridge, MA: National Bureau of Economic Research. Autor, D. (2013). The “task approach” to labor markets an overview (No. w18711). Cambridge, MA: National Bureau of Economic Research. Autor, D. H., & Dorn, D. (2013). The growth of low-skill service jobs and the polarization of the US labor market. American Economic Review, 103(5), 1553–1597. doi:10.1257/ aer.103.5.1553 Autor, D. H., Katz, L. F., & Kearney, M. S. (2006). The polarization of the US labor market. American Economic Review, 96(2), 189–194. doi:10.1257/000282806777212620 Autor, D. H., Levy, F., & Murnane, R. J. (2003). The skill content of recent technological change: An empirical exploration. The Quarterly Journal of Economics, 118(4), 1279– 1333. Avery, C., Resnick, P., & Zeckhauser, R. (1999). The market for evaluations. American Economic Review, 89(3), 564–584. doi:10.1257/aer.89.3.564 Ba, S. (2001). Establishing trust through a community responsibility system. Decision Support Systems, 31(3), 323–336. Badger, E. (2014, March 25). Why Airbnb wants you to know how much its users are spending. The Washington Post. Retrieved from http://www.washingtonpost.com/blogs/ wonkblog/wp/2014/03/25/why-airbnb-wants-you-to-know-how-much-its-users-arespending/. Accessed on May 28, 2015. Bailey, K. D. (1994). Typologies and taxonomies: An introduction to classification techniques. Thousand Oaks, CA: SAGE. Bajari, P., & Hortasu, A. (2004). Economic insights from internet auctions. Journal of Economic Literature, 42(2), 457–486. doi:10.1257/0022051041409075

References    171 Baldry, C. (2011). Editorial: Chronicling the information revolution. New Technology, Work and Employment, 26(3), 175–182. doi:10.1111/j.1468-005X.2011.00267.x Banning, M. E. (2016). Shared entanglements: Web 2.0, info-liberalism & digital sharing. Information, Communication & Society, 19(4), 489–503. doi:10.1080/1369118X. 2015.1061573 Baranowska, A., & Gebel, M. (2010). The determinants of youth temporary employment in the enlarged Europe. European Societies, 12(3), 367–390. doi:10.1080/14616690903165434 Barbezieux, P., & Herody, C. (2016). Rapport Au Premier Ministre Sur L’économie collaborative. Paris: République Français, Premier Ministre. Retrieved from http://www. gouvernement.fr/sites/default/files/liseuse/6421/master/projet/Rapport-de-PascalTerrasse-sur-le-d%C3%A9veloppement-de-l-%C3%A9conomie-collaborative.pdf. Accessed on February 11, 2016. Bardhi, F., & Eckhardt, M. (2012). Access-based consumption: The case of car sharing. Journal of Consumer Research, 39(4), 881–898. Bar-Isaac, H., Caruana, G., & Cuñat, V. (2012). Information gathering externalities for a multi-attribute good. Journal of Industrial Economics, 60(1), 162–185. Barnes, S. J., & Mattsson, J. (2015). Understanding current and future issues in collaborative consumption: A four-stage Delphi study. Technological Forecasting and Social Change. 104, 200–211. doi:10.1016/j.techfore.2016.01.006 Barry, J., & Caron, P. (2014). Tax regulation, transportation innovation, and the sharing economy. University of Chicago Law Review Dialogue, 82, 2015; Pepperdine University Legal Studies Research Paper No. 2015/2, San Diego Legal Studies Paper No. 15-180. Retrieved from http://ssrn.com/abstract=2538947 or http://dx.doi.org/10.2139/ ssrn.2538947 Baumeister, C., & Wangenheim, F. (2014). Access vs. ownership: Understanding consumers’ consumption mode preference. Retrieved from http://ssrn.com/abstract=2463076. Accessed on August 15, 2015. Baumgartner, F. R., Breunig, C., Green-Pedersen, C., Jones, B. D., Mortensen, P. B., Nuytemans, M. & Walgrave, S. (2009). Punctuated equilibrium in comparative perspective. American Journal of Political Science, 53(3), 603–620. doi:10.1111/j.1540– 5907.2009.00389.x Beerepoot, N., & Lambregts, B. (2015). Competition in online job marketplaces: Towards a global labour market for outsourcing services? Global Networks, 15(2), 236–255. doi:10.1111/glob.12051 Béland, D. (2009). Ideas, institutions, and policy change. Journal of European Public Policy, 16(5), 701–718. doi:10.1080/13501760902983382 Belk, R. (2014a). Sharing versus pseudo-sharing in web 2.0. Anthropologist, 18(1), 7–23. Belk, R. (2014b). You are what you can access: Sharing and collaborative consumption online. Journal of Business Research, 67(8), 1596–1600. doi:10.1016/j.jbusres.2013.10.001 Benford, R., & Snow, D. (2000). Framing processes and social movements: An overview and assessment. Annual Review of Sociology, 26, 611–639. Benjaafar, S., Kong, G., Li, X., & Courcoubetis, C. (2015). Peer-to-peer product sharing: Implications for ownership, usage and social welfare in the sharing economy. Retrieved from http://ssrn.com/abstract=2669823. Accessed on March 28, 2016. Benkler, Y. (2004). Sharing nicely: On shareable goods and the emergence of sharing as a modality of economic production. Yale Law Journal, 114(2), 273–358. Benkler, Y. (2006). The wealth of networks: How social production transforms markets and freedom. New Haven, CT: Yale University Press. Benkler, Y., & Nissenbaum, H. (2006). Commons-based peer production and virtue. The Journal of Political Philosophy, 14(4), 394–419. Bennett, W. L. (2004). Communicating global activism. In W. Van de Donk, B. Loader, P. Nixon, & D. Rucht (Eds.), Cyberprotest: New media, citizens and social movements. (pp. 109–128) London: Routledge.

172   References Bensinger, G. (2015a, August 6). Sprig Is the latest startup to hire its contract workers. The Wall Street Journal. Retrieved from http://blogs.wsj.com/digits/2015/08/06/sprig-islatest-startup-to-hire-its-contract-workers/. Accessed on January 21, 2016. Bensinger, G. (2015b, July 30). Start-ups scramble to define “employee.” The Wall Street Journal. Retrieved from http://www.wsj.com/articles/startups-scramble-to-defineemployee-1438228860. Accessed on January 21, 2016. Benson, A., Sojourner, A., & Umyarov, A. (2015). Can reputation discipline the gig economy? Experimental evidence from an online labor market. IZA Discussion Paper No. 9501. Retrieved from https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2696299 Berg, J. (2016). Income security in the on-demand economy: Findings and policy lessons from a survey of crowdworkers. Comparative Labor Law & Policy Journal, 37(3). Retrieved from http://ssrn.com/abstract=2740940. Accessed on May 2, 2016. Bergvall-Kåreborn, B., & Howcroft, D. (2014). Amazon mechanical turk and the commodification of labour. New Technology, Work and Employment, 29(3), 213–223. doi:10.1111/ntwe.12038 Bernhardt, A. (2014). Labor standards and the reorganization of work: Gaps in data and research. IRLE Working Paper #100-14. Berkeley. Retrieved from http://www.irle. berkeley.edu/workingpapers/100-14.pdf Bernstein, M. S., Little, G., Miller, R. C., Hartman, B., Ackerman, M. S., Karger, D. R., Crowell, D., & Panovich, K. (2015). Soylent: A word processor with a crowd inside. Communications of the ACM, 58(8), 85–94. doi:10.1145/2791285 Beyers, J., Eising, R., & Maloney, W. (2008). Researching interest group politics in Europe and elsewhere: Much we study, little we know? West European Politics, 31(6), 1103–1128. doi:10.1080/01402380802370443 Bihr, A. (1989). Entre Bourgeoisie et Prolétariat: L’encadrement Capitaliste, Logiques sociales, Paris: L’Harmattan. Bilis, M. (2015, June 22). Instacart converts some independent contractors into parttime workers. San Francisco Business Times. Retrieved from http://www.bizjournals. com/sanfrancisco/blog/2015/06/instacart-indepedent-contractors-workers-uber.html. Accessed on January 21, 2016. Bogliacino, F., Codagnone, C., & Veltri, G. (2015). The behavioural turn in consumer policy: Perspectives and clarifications. Intereconomics, 50(2), 108–114. doi:10.1007/s10272015-0532-4 Boltanski, L., & Thévenot, L. (2006). On justification: Economies of worth (C. Porter Trans.). Princeton, NJ: Princeton University Press, Bolton, G., Greiner, B., & Ockenfels, A. (2012). Engineering trust: Reciprocity in the production of reputation information. Management Science, 59(2), 265–285. doi:10.1287/mnsc.1120.1609 Bond, A. (2015). An app for that: Local governments and the rise of the sharing economy. Notre Dame Law Review, 90(2), 77–96. Botsman, R., & Rogers, R. (2010). What’s mine is yours: How collaborative consumption is changing the way we live. New York, NY: Harper Business. Bouquillion, P., & Matthews, J. (2010). Le web collaboratif, mutations des industries de la culture et de la communication. Grenoble, France: Pug. Bourdieu, P. (1984). Distinction: A social critique of the judgement of taste. Cambridge, MA: Harvard University Press. Bourdieu, P. (1986). The forms of capital. In J. Richardson (Ed.), Handbook of theory and research for the sociology of education (pp. 241–258). New York, NY: Greenwood. Bouwen, P. (2002). Corporate lobbying in the European Union: The logic of access. Journal of European Public Policy, 9(3), 365–390. Bouwen, P. (2004). Exchanging access goods for access: A comparative study of business lobbying in the European Union institutions. European Journal of Political Research, 43(3), 337–369. doi:10.1111/j.1475-6765.2004.00157.x Bowles, J. (2014, July 24). The computerisation of European jobs. Retrieved from http://bruegel. org/2014/07/the-computerisation-of-european-jobs/. Accessed on December 28, 2015.

References    173 Brabham, D. C. (2008). Crowdsourcing as a model for problem-solving: An introduction and cases. Convergence: The International Journal of Research into New Media Technologies, 14(1), 75–90. Brabham, D. (2013). Crowdsourcing. Cambridge, MA: MIT Press. Bradshaw, D. (2015, 22 April). Sharing economy benefits lower income groups. Financial Times. Retrieved from http://www.ft.com/cms/s/2/7afde9b0-d95a-11e4-a8f100144feab7de.html-axzz3bkTK3f8R. Accessed on May 31, 2015. Brevini, B., Hintz, A., & McCurdy, P. (Eds.). (2013). Beyond Wikileaks: Implications for the future of communications, journalism and society. New York, NY: Palgrave Macmillan. Brewster, C., Mayne, L., & Tregaskis, O. (1997). Flexible staffing in Europe. Journal of World Business, 32(2), 133–151. Brown, A. (2015). Taskrabbit workers can take home $7,000 a month for doing other people’s chores. Retrieved from http://madamenoire.com/520031/taskrabbit-workers-can-takehome-7000-a-month-for-doing-other-peoples-chores/. Accessed on November 30, 2015. Brownsword, R., & Goodwin, M. (2012). Law and the technologies of the twenty-first century. Cambridge: Cambridge University Press. Brownsword, R., & Somsen, H. (2009). Law, innovation and technology: Before we fast forward – A forum for debate. Law, Innovation and Technology, 1(1), 1–73. doi:10.1080/ 17579961.2009.11428364 Bruns, A. (2007). Produsage: Towards a Broader Framework for User-Led Content Creation. In Proceedings Creativity & Cognition (p. 6), Washington, DC. http://eprints. qut.edu.au/6623/1/6623.pdf Brynjolfsson, E., & McAfee, A. (2012). Race against the machine: How the digital revolution is acclerating innovation, driving productivity and irreversibly transforming employment and the economy. Mexington, MA: Digital Frontier Press. Brynjolfsson, E., & McAfee, A. (2014). The second machine age: Work, progress, and prosperity in a time of brilliant technologies (1st ed.). New York, NY: W. W. Norton. Bucher, E., Fieseler, C., & Lutz, C. (2016). What’s mine is yours (for a nominal fee): Exploring the spectrum of utilitarian to altruistic motives for Internet-mediated sharing. Computers in Human Behavior, 62, 316–326. doi:https://doi.org/10.1016/j.chb.2016.04.002 Burchell, B. (2009). Flexicurity as a moderator of the relationship between job insecurity and psychological well-being. Cambridge Journal of Regions, Economy and Society, 2(3), 365–378. doi:10.1093/cjres/rsp021 Burgoon, B. & Dekker, F. (2010). Flexible employment, economic insecurity and social policy preferences in Europe . Journal of European Social Policies, 20(2), 126–141. Burnham, S. (2011). The publicization of trust. Privatize, 30, 3–7. Buschoff, K. S., & Protsch, P. (2008). (A-)typical and (in-)secure? Social protection and “non-standard” forms of employment in Europe. International Social Security Review, 61(4), 51–73. doi:10.1111/j.1468-246X.2008.00323.x Butenko, A., & Larouche, P. (2015). Regulation for innovativeness or regulation of innovation? Law, Innovation and Technology, 7(1), 52–82. doi:10.1080/17579961.2015.1052643 Cabinet Office. (1999). Modernising government. London: Stationery Office. Cabral, L., & Hortacsu, A. (2010). The dynamics of seller reputation: Evidence from eBAY. The Journal of Industrial Economics, 58(1), 54–78. doi:10.1111/j.1467-6451.2010.00405.x Cahn, E., & Gray, C. (2015). The time bank solution. Stanford Social Innovation Review, 13(3), 40–45. Cahn, E., & Rowe, J. (1992). Time dollars. Emmaus, PA: Rodale Press. Cahuc, P., & Kramarz, F. (2005). De la précarité à mobilité, vers une sécurité sociale professionnelle: Rapport au ministre de l’économie, des finances et de l’industrie et au ministre de l’emploi, du travail et de la cohésion sociale. Paris, France: La Documentation Francaise. Caillaud, B., & Jullien, B. (2003a). Chicken & egg: Competition among intermediation service providers. RAND Journal of Economics, 34(2), 309–328. Caillaud, B., & Jullien, B. (2003b). Chicken & egg: Competition among intermediation service providers. The RAND Journal of Economics, 34(2), 309–328.

174   References Caldararo, N. L. (2014). Empty housing in America, housing needs, shadow banks and the sharing economy. Retrieved from http://ssrn.com/abstract=2460987. Accessed on July 26, 2015. Cannon, B., & Chung, H. (2015). A framework for designing co-regulation models well-adapted to technology-facilitated sharing economies. Santa Clara High Tech. L. J., 31, 23–97. Cannon, S., & Summers, L. (2014, October). How Uber and the sharing economy can win over regulators. Harvard Business Review. Retrieved from https://hbr.org/2014/10/howuber-and-the-sharing-economy-can-win-over-regulators/ Cardon, D., & Casili, A. (2015). Qu’est-ce que le digital labor? Paris, France: Editions de l’INA. Carr, N. (2008). The big switch: Rewiring the world, from Edison to Google. New York, NY: W. W. Norton & Company. Castells, M. (2000). The information age: Economy, society and culture (2nd ed.). Vol.1: The rise of the network society. Oxford: Blackwell. Castillo, J., Knoepfle, D., & Weyl, E. (2018). Surge pricing solves the wild goose chase. Retrieved from https://ssrn.com/abstract=2890666. Accessed on May 2018. CEPR. (2015, January 23). Ubernomics. Retrieved from http://journalistsresource.org/ studies/economics/business/airbnb-lyft-uber-bike-share-sharing-economy-researchroundup. Accessed on June 10, 2015. Chadwick, A. (2006) Internet politics: States, citizens, and new communication technologies. Oxford: Oxford University Press. Chandler, D., & Kapelner, A. (2013). Breaking monotony with meaning: Motivation in crowdsourcing markets. Journal of Economic Behavior and Organisation, 90, 123–133. Chen, M., Chevalier, J., Rossi, P., & Oehlsen, E. (2017). The value of flexible work: Evidence from Uber drivers. NBER Working Paper Series No. 23296 (p. 19525). Retrieved from http://www.nber.org/papers/w23296. Accessed on May 2018. Chen, L., Mislove, A., & Wilson, C. (2015). Peeking beneath the hood of Uber. Paper presented at the Proceedings of the 2015 ACM conference on internet measurement conference, Tokyo, Japan. Cherry, M. (2011). A taxonomy of virtual work. Georgia Law Review, 45, 951–990. Cherry, M. (2016). Beyond misclassification: The digital transformation of work. Comparative Labor Law & Policy Journal, 37(3). Retrieved from http://ssrn.com/ abstract=2734288. Accessed on May 3, 2016. Cherry, M., & Aloisi, A. (2016). Dependent contractors in the gig economy: A comparative approach. American University Law Review, 66, 635–689. CIPD. (2017). To gig or not to gig? Stories from the modern economy. London: Chartered Institute of Personnel and Development. Cockayne, D. G. (2016). Sharing and neoliberal discourse: The economic function of sharing in the digital on-demand economy. Geoforum, 77, 73–82. Retrieved from https:// doi.org/10.1016/j.geoforum.2016.10.005 Codagnone, C. (2017). Lobbying as rhetorical framing in the “sharing economy”: A case study on the limits and crisis of the evidence based policy paradigm. DigiWorld Economic Journal, 108, 15–42. Codagnone, C., Abadie, F., & Biagi, F. (2016a). The future of work in the sharing economy: Market efficiency and equitable opportunities or unfair precarisation? JRC Science and Policy Report. Joint Research Centre, Institute for Prospective Technological Studies, European Commission. Sevilla. Codagnone, C., Abadie, F., & Biagi, F. (2016b). The passions and the interests: Unpacking the sharing economy. JRC Science and Policy Report, Joint Research Centre, Institute for Prospective Technological Studies, European Commission. Codagnone, C., Bogliacino, F., & Veltri, G. A. (2018a). Scienza in vendita. Incertezza, interessi e valori nelle politiche pubbliche. Milan, Italy: Egea. Codagnone, C., Lupiáñez-Villanueva, F., Tornese, P. Gaskell, G., Veltri, G. A., Villa, J., Franco, Y., Vitiello, S., Theben, A., Ortoleva, P., Cirillo, V., & Fana, M. (2018b). Behavioural study on the effects of an extension of access to social protection for people in

References    175 all forms of employment. Brussels, Belgium: European Commission, DG Employment Social Affairs and Inclusion. Codagnone, C., & Martens, B. (2016c). Scoping the sharing economy: Origins, definitions, impact and regulatory issues. JRC Technical Report, Digital Economy Working Paper 2016/01. Sevilla, Spain: Joint Research Centre, Institute for Prospective Technological Studies, European Commission. Cohen, P., Hahn, R., Hall, J., Levitt, S., & Metcalfe, R. (2016). Using big data to estimate consumer surplus: The case of Uber. NBER Working Paper Series No. 22627 (p. 19525). Retrieved from http://www.nber.org/papers/w22627. Accessed on May 2018. Cohen, B., & Muñoz, P. (2015). Sharing cities and sustainable consumption and production: Towards an integrated framework. Journal of Cleaner Production. doi:10.1016/ j.jclepro.2015.07.133 Cohen, M., & Sundararajan, A. (2015). Self-regulation and innovation in the peer-to-peer sharing economy. The University of Chicago Law Review Dialogue, 82, 116–133. Cohen, M., & Zehngebot, C. (2014). What’s old becomes new: Regulating the sharing economy. Boston Bar Journal, 34. Retrieved from http://bostonbarjournal. com/2014/2004/2001/whats-old-becomes-new-regulating-the-sharing-economy/ Coleman, J. (1988). Social capital in the creation of human capital. The American Journal of Sociology, 94, S95. Coleman, J. S. C. (1990). Foundations of social theory. Cambridge MA: The Belknap Press of the Harvard University Press. Collier, D., LaPorte, J., & Seawright, J. (2012). Putting typologies to work: Concept formation, measurement, and analytic rigor. Political Research Quarterly, 65(1), 217–232. Collins, K. (2014, May 30). Uber welcomes TfL’s taximeter High Court referral. Wired. Retrieved from http://www.wired.co.uk/news/archive/2014-05/30/uber-tfl-high-court. Accessed on June 8, 2015. Collom, E., Lasker, J., & Kyriacou, C. (2012). Equal time, equal value: Community currencies and time banking in the US. Farnham: Ashgate Publishing. Comor, E. (2010). Digital prosumption and alienation. Ephemera, 10(3/4): 439–454. Cook, C., Diamond, R., Hall, J., List, J., & Oyer, P. (2018). The gender earnings gap in the gig economy: Evidence from over a million rideshare drivers. Retrieved from https://web. stanford.edu/~diamondr/UberPayGap.pdf. Accessed on May 2018. Corciolani, M., & Dalli, D. (2014). Gift-giving, sharing and commodity exchange at Bookcrossing.com: New insights from a qualitative analysis. Management Decision, 52(4), 755–755. Corritore, C. L., Kracher, B., & Wiedenbeck, S. (2003). On-line trust: Concepts, evolving themes, a model. International Journal of Human–Computer Studies, 58(6), 737–758. doi:10.1016/S1071-5819(03)00041-7 Couch, L., & Jones, W. (1997). Measuring levels of trust. Journal of Research in Personality, 31(3), 319–336. Cramer, J., & Krueger, A. (2016). Disruptive change in the taxi business: The case of uber. American Economic Review, 106(5), 177–182. Cruciatti, P. (2018, April 12). “Cosa ci dice la sentenza di ieri su Foodora.” Il Post. Retrieved from https://www.ilpost.it/2018/04/12/foodora-sentenza/. Accessed on May 2018. Cullen, Z., & Farronato, C. (2015). Outsourcing tasks online: Matching supply and demand on peer-to-peer internet platforms. Stanford, CA: Department of Economics, Stanford University. Retrieved from http://web.stanford.edu/~chiaraf/matching_p2p_latest.pdf Dahlberg, L., & Siapera, E. (2007). Radical politics and the internet. Basingstoke: Palgrave Macmillan. Damiano, E., & Li, H. (2008). Competing matchmaking. Journal of the European Economic Association, 6(4), 789–818. Daus, M., & Russo, P. (2015). One standard for all. Criminal background checks for taxicab, for-hire, and transportation network company (Tnc) drivers. New York, NY: Jay College of Criminal Justice of the City University of New York. Retrieved from

176   References http://www.utrc2.org/sites/default/files/pubs/Background Check Report.pdf. Accessed on September 7, 2015. Deakin, S. (2014). New forms of employment: Implications for EU-law – The law as it stands. Paper presented at the European Labour Law Network – 7th Annual Legal Seminar: New Forms of Employment and EU Law, The Hague, The Netherlands. Retrieved from http:// www.fd.unl.pt/docentes_docs/ma/jja_MA_24419.pdf. Accessed on January 28, 2016. DeAmicis, C. (2015, July 17). Homejoy shuts down after battling worker classification lawsuits. Recode. Retrieved from http://recode.net/2015/07/17/cleaning-services-startuphomejoy-shuts-down-after-battling-worker-classification-lawsuits/. Accessed on November 30, 2015. de Graaf-Zijl, M., van den Berg, G. J., & Heyma, A. (2011). Stepping stones for the unemployed: the effect of temporary jobs on the duration until (regular) work. Journal of Population Economics, 24(1), 107–139. doi:10.1007/s00148-009-0287-y De Groen, W., Kilhoffer, Z., Lenaerts, K., & Salez, N. (2017). The impact of the platform economy on job creation. Intereconomics, 52(6), 345–351. Degryse, C. (2016). Digitalisation of the economy and its impact on labour markets. ETUI Working Paper 2. Brussels, Belgium: ETUI. Delhey, J., Newton, K., & Welzel, C. (2011). How general is trust in “most people”? Solving the radius of trust problem. American Sociological Review, 76(5), 786–807. Dellarocas, C., & Wood, C. A. (2007). The sound of silence in online feedback: Estimating trading risks in the presence of reporting bias. Management Science, 54(3), 460–476. doi:10.1287/mnsc.1070.0747 Demailly, D., & Novel, A.-S. (2014). The sharing economy: Make it sustainable. Paris, France: IDDRI, SciencePo. Retrieved from http://www.iddri.org/Publications/ Collections/Analyses/ST0314_DD ASN_sharing economy.pdf Dencik, L., Hintz, A., & Cable, J. (2016). Toward data justice? The ambiguity of antisurveillance resistance in political activism. Big Data and Society, 3(2). Retrieved from https://doi.org/10.1177/2053951716679678 Department for Business, Energy and Industrial Strategy (BEIS). (2018). The experiences of individuals in the gig economy. London: BEIS. De Stefano, V. (2016). The rise of the just-in-time workforce: On-demand work, crowdwork and labour protection in the gig-economy. Conditions of Work and Employment Series No. 71. Geneva: ILO, Inclusive Labour Markets, Labour Relations and Working Conditions Branch. Dettling, L. (2016). Broadband in the labor market: The impact of residential high speed internet on married women’s labor force participation. Industrial and Labor Relations Review. doi:10.1177/0019793916644721. Retrieved from http://ilr.sagepub.com/content/ early/2016/04/20/0019793916644721.abstract Deuze, M. (2007). Mediawork. Cambridge, MA: Polity Press. Diani, M., & McAdam, D. (2003). Social movements and networks: Relational approaches to collective action. Oxford: Oxford University Press. Dillahunt, T., & Malone, A. (2015). The promise of the sharing economy among disadvantaged communities. Paper presented at the Proceedings of the 33rd annual ACM conference on human factors in computing systems, Seoul, Republic of Korea. DiMaggio, P., & Powell, W. (1983). The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review, 48, 147–160. DiMaggio, P., & Powell, W. (1991). The new institutionalism in organizational analysis. Chicago, IL: University of Chicago Press. Dinerstein, M., Einav, L., Levin, J., & Sundaresan, N. (2014). Consumer price search and platform design in internet commerce. National Bureau of Economic Research Working Paper Series, No. 20415. doi:10.3386/w20415 DiPrete, T. A., Goux, D., Maurin, E., & Quesnel-Vallee, A. (2006). Work and pay in flexible and regulated labor markets: A generalized perspective on institutional evolution and

References    177 inequality trends in Europe and the U.S. Research in Social Stratification and Mobility, 24(3), 311–332. doi:10.1016/j.rssm.2006.04.001 Djelassi, S., & Decoopman, I. (2013). Customers’ participation in product development through crowdsourcing: Issues and implications. Industrial Marketing Management, 42(5), 683–692. doi:http://dx.doi.org/10.1016/j.indmarman.2013.05.006 Dobbs, R., Remes, J., Manyika, J., Roxburgh, C., Smit, S., & Schaer, F. (2012). Urban world: Cities and the rise of the consuming class. Washington, DC: McKinsey Global Institute. Retrieved from http://www.mckinsey.com/~/media/McKinsey/dotcom/Insights and pubs/ MGI/Research/Urbanization/Urban world - Rise of the consuming class/MGI_Urban_ world_Rise_of_the_consuming_class_Full_report.ashx. Accessed on August 19, 2015. Dobbs, R., Smit, S., Remes, J., Manyika, J., Roxburgh, C. & Restrepo, A. (2011). Urban world: Mapping the economic power of cities. Washington, DC: McKinsey Global Institute. Retrieved from http://www.mckinsey.com/~/media/McKinsey/dotcom/ Insights and pubs/MGI/Research/Urbanization/Urban world mapping economic power of cities/MGI_urban_world_mapping_economic_power_of_cities_full_report. ashx. Accessed on August 19, 2015. Dokko, J., Mumford, M., & Schanzenbach, D. (2015). Workers and the online gig economy. A Hamilton Project Framing Paper. Washington, DC: The Hamilton Project, Brookings. Dolata, U. (2017a). Social movements and the internet: The sociotechnical constitution of collective action. Discussion Paper No. 2017-02. Stuttgart, Germany: Stuttgarter Beiträge zur Organisations-und Innovationsforschung, SOI. Dolata, U. (2017b). Apple, Amazon, Google, Facebook, Microsoft: Market concentration– competition–innovation strategies. SOI Discussion Paper No. 2017-01. Stuttgart, Germany: Stuttgarter Beiträge zur Organisations und Innovationsforschung. Dolata, U., & Schrape, J.-F. (2016). Masses, crowds, communities, movements. Collective action in the internet age. Social Movement Studies, 15(1), 1–18. Dowling, E. (2011). Challenging the Critical Subject: Pedagogies of Cognitive Capitalism. In Michael, A. P., & Ergin, B. (eds.), Cognitive Capitalism, Education, and Digital Labor, New York: Peter Lang. Downs, J., Holbrook, M., Sheng, S., & Cranor, L. (2010). Are your participants gaming the system? Screening mechanical turk workers. Paper presented at the Proceedings of the SIGCHI conference on human factors in computing systems, Atlanta, GA. Dredge, D., & Gyimóthy, S. (2015). The collaborative economy and tourism: Critical perspectives, questionable claims and silenced voices. Tourism Recreation Research, 40(3), 286–302. doi:10.1080/02508281.2015.1086076 Drucker, P. (1993) The post-capitalist society. Oxford: Buttherworth-Heinemann. Drutman, L. (2016). The business of America is lobbying: How corporations became politicized and politics became more corporate. Oxford: Oxford University Press. Dubal, V. (2017). Wage slave or entrepreneur? Contesting the dualism of legal worker identities. California Law Review, 105(1), 65–126. Dubois, E., Schor, J., & Carfagna, L. (2014). New cultures of connection in a Boston Time Bank. In J. Schor & C. Thompson (Eds.), Practicing plenitude (pp. 95–114). New Haven, CT: Yale University Press. Edelman, B., & Geradin, D. (2016). Efficiencies and regulatory shortcuts: How should we regulate companies like Airbnb and Uber? 19(2), 293–328. Edelman, B., & Luca, M. (2014). Digital discrimination: The case of Airbnb, Boston, MA: Harvard Business School Working Papers. Eichhorst, W. (2013). The unequal distribution of labor market risks: Permanent vs. temporary employment. Bonn, Germany: IZA VEF. Eichhorst, W., Braga, M., Famira-Mühlberger, U., Gerard, M., Horvath, T., Kahanec, M., Kahancová, M., Kendzia, M., Martišková, M., Monti, P., Pedersen, J., Stanley, J., Vandeweghe, B., Wehner, C., & White, C. (2013). Social protection rights of economically dependent self-employed workers. Brussels, Belgium: European Parliament.

178   References Einav, L., & Levin, J. (2014). The data revolution and economic analysis. Innovation Policy and the Economy, 14(1), 1–24. doi: 10.1086/674019 Einav, L., Farronato, C., & Levin, J. (2015). Peer-to-peer markets. National Bureau of Economic Research Working Paper Series, No. 21496. doi:10.3386/w21496 Einav, L., Farronato, C., Levin, J. D., & Sundaresan, N. (2013). Sales mechanisms in online markets: What happened to internet auctions? National Bureau of Economic Research Working Paper Series, No. 19021. doi:10.3386/w19021 Eisenmann, T., Parker, G., & Van Alstyne, M. (2006). Strategies for two-sided markets. Harvard Business Review, 84(10), 92–101. Elance-oDesk. (2014). Annual impact report. Retrieved from https://blog-static.odesk.com/ content/Elance-oDeskAnnualImpactReport2014.pdf. Accessed on May 28, 2015. Emmenegger, P., Palier, B., & Seeleib-Kaiser, M. (2012). The age of dualization. The changing face of inequality in deindustrializing societies. Oxford: Oxford University Press. Ert, E., Fleischer, A., & Magen, N. (2016). Trust and reputation in the sharing economy: The role of personal photos in Airbnb. Tourism Management, 55, 62–73. doi:https://doi. org/10.1016/j.tourman.2016.01.013 Esping-Andersen, G. (1990). The three worlds of welfare capitalism. Princeton, NJ: Princeton University Press. Esping-Andersen, G., & Regini, M. (Eds.). (2000). Why deregulate labour markets? New York, NY: Oxford University Press. Eurofound. (2015). New forms of employment. Luxembourg: Publications Office of the European Union. European Commission. (2013). Employment and social developments in Europe 2013. Luxembourg: Publications Office of the European Union. European Commission. (2014a). A decade of labour market reforms in the EU: Insights from the LABREF database. Economic Papers, Brussels, Belgium: European Commission. European Commission. (2014b). Employment and social developments in Europe 2014. Luxembourg: Publications Office of the European Union. European Commission. (2015a). A digital single market strategy for Europe – Analysis and evidence. Accompanying the communication a digital single market strategy for Europe SWD (2015) 100 final. Commission Staff Working Document. Brussels, Belgium: European Commission. European Commission. (2015b). A digital single market strategy for Europe. COM (2015) 192 Final Report. Brussels, Belgium: European Commission. European Commission. (2015c). Employment and social developments in Europe 2015. Luxembourg: Publications Office of the European Union. European Commission. (2016a). Employment and social developments in Europe 2016. Luxembourg: Publications Office of the European Union. European Commission. (2016b). A European agenda for the collaborative economy. COM (2016) 356 Final Report. Brussels, Belgium: European Commission. European Commission. (2016c). Online platforms and the digital single market. Opportunities and challenges for Europe. COM (2016) 288 Final Report. Brussels, Belgium: European Commission. European Commission. (2016d). The use of collaborative platforms Flash Eurobarometer 438. Brussels, Belgium: European Commission. European Commission. (2018). Proposal for a council recommendation on access to social protection for workers and the self-employed. COM (2018) 132 Final Report. Brussels, Belgium: European Commission. European Economic and Social Committee (EESC). (2014). Opinion of the European Economic and Social Committee on collaborative or participatory consumption, a sustainability model for the 21st century. Brussels, Belgium: European Economic and Social Committee (EESC), INT/686. European Parliament. (2016). The cost of non-Europe in the sharing economy. Brussels, Belgium: European Parliament, European Parliamentary Research Service.

References    179 European Parliament. (2017a). Online platforms: How to adapt regulatory framework to the digital age? Briefing, No. 607.323, Brussels, Belgium and Strasbourg: European Parliament. European Parliament. (2017b). The social protection of workers in the platform economy. Brussels, Belgium: European Parliament, Directorate General for Internal Policy. European Trade Union Institute (ETUI). (2012). Benchmarking working Europe 2012. Brussels, Belgium: ETUI. European Trade Union Institute (ETUI). (2015). Benchmarking working Europe 2015. Brussels, Belgium: ETUI. Evans, D. (2003a). The antitrust economics of multi-sided platform markets. Yale Journal on Regulation, 20(2), 325–381. Evans, D. (2003b). Some empirical aspects of multi-sided platform industries. Review of Network Economics, 2(3), 2194–5993. doi:10.2202/1446-9022.1026 Evans, D. (2008a). Competition and regulatory policy for multi-sided platforms with applications to the web economy. SSRN Electronic Journal2(March). doi:10.2139/ ssrn.1090368 Evans, D. (2008b). The economics of the online advertising industry. Review of Network Economics, 7(3), 351–391. doi:10.2202/1446-9022.1154 Evans, D. (2009). The online advertising industry: Economics, evolution, and privacy. Journal of Economic Perspectives, 23(3), 37–60. doi:10.1257/jep.23.3.37 Evans, D. (2011). Platforms economics: Essays on multi-sided businesses. Competition Policy International. Evans, P., & Gawer, A. (2016). The rise of the platform enterprise. A Global Survey. New York, NY: Center for Global Enterprise. Evans, D., & Noel, M. (2005). Defining antitrust markets when firms operate two-sided platforms. Columbia Business Law Review, 3, 667–702. Evans, D., & Schmalensee, R. (2007). The industrial organization of markets with twosided platforms. Competition Policy International, 3(1), 151–179. Fabo, B., Beblavý, M., Kilhoffer, Z., & Lenaerts, K. (2017). Overview of European platforms: Scope and business models JRC Science for Policy Report. Seville: European Commission, Joint Research Centre Institute for Prospective Technological Studies. Fang, B., Ye, Q., & Law, R. (2015). Effect of sharing economy on tourism industry employment. Annals of Tourism Research. 57, 264–267. doi:10.1016/j.annals.2015.11.018 Farronato, C., & Fradkin, A. (2015). Market structure with the entry of peer-to-peer platforms: The case of hotels and Airbnb. Stanford, CA: Stanford University. Felstiner, A. (2011). Working the crowd: Employment and labor law in the crowdsourcing industry. Berkeley Journal of Employment and Labor Law, 32, 143–203. Ferber, M., & Waldfogel, J. (1998). The long-term consequences of non-traditional employment. Monthly Labor Review, 121(5), 3–12. Fischer, E. (1996). How to read Karl Marx. New York, NY: Monthly Review Press. Filistrucchi, L., Geradin, D., & Van Damme, E. (2013). Identifying two-sided markets. World Competition: Law & Economics Review, 36(1), 33–60. Filistrucchi, L., Geradin, D., van Damme, E., & Affeldt, P. (2014). Market definition in two-sided markets: Theory and practices. Journal of Competition Law and Economics, 10(2), 293–339. doi:10.1093/joclec/nhu007 Fitzmaurice, C. J., Ladegaard, I., Attwood-Charles, W., Cansoy, M., Carfagna, L. B., Schor, J. B., & Wengronowitz, R. (2018). Domesticating the market: Moral exchange and the sharing economy. Socio-Economic Review. mwy003, https://doi.org/10.1093/ser/ mwy003 Article Florisson, R., & Mandl, I. (2018). Platform work: Types and implication for work and employment – Literature review. Dublin: Eurofound. Forno, F., & Garibaldi, R. (2015). Sharing economy in travel and tourism: The case of home-swapping in Italy. Journal of Quality Assurance in Hospitality & Tourism, 16(2), 202–220. doi:10.1080/1528008X.2015.1013409

180   References Fradkin, A. (2014). Search frictions and the design of online marketplaces. Retrieved from http://andreyfradkin.com/assets/Fradkin_JMP_Sep2014.pdf Fradkin, A., Grewal, E., Holtz, D., & Pearson, M. (2015). Bias and reciprocity in online reviews: Evidence from field experiments on Airbnb. In Proceedings of the Sixteenth ACM Conference on Economics and Computation, Portland, OR, June 15–19, 2015 (p. 641). New York, NY: ACM. Fraiberger, S., & Sundararajan, A. (2015). Peer-to-peer rental markets in the sharing economy. NYU Stern School of Business Research Paper, New York. Retrieved from http:// dx.doi.org/10.2139/ssrn Freelancers Union & Elance-oDesk. (2014). Freelancing in America: A national survey of the new workforce. An independent study commissioned by Freelancers Union & ElanceoDesk. Retrieved from https://fu-web-storage-prod.s3.amazonaws.com/content/filer_ public/7c/45/7c457488-0740-4bc4-ae45-0aa60daac531/freelancinginamerica_report. pdf. Accessed on June 9, 2015. Freelancers Union & Upwork. (2015). Freelancing in America 2015. An independent study commissioned by Freelancers Union & Upwork. Retrieved from https://fu-webstorage-prod.s3.amazonaws.com/content/filer_public/59/e7/59e70be1-5730-4db8-919f1d9b5024f939/survey_2015.pdf. Accessed on December 2, 2015. Freeman, R. (2008). The new global labor market. Focus, 26(1), 1–6. Freitag, M., & Traunmüller, R. (2009). Spheres of trust: An empirical analysis of the foundations of particularised and generalised trust. European Journal of Political Research, 48(6), 782–803. doi:10.1111/j.1475-6765.2009.00849.x Frenken, K., Meelen, T., Arets, M., & van de Glind, P. (2015, 20 May). Smarter regulation for the sharing economy, The Guardian. Retrieved from http://www.theguardian.com/ science/political-science/2015/may/20/smarter-regulation-for-the-sharing-economy. Accessed on May 31, 2015. Frenken, K., & Schor, J. (2017). Putting the sharing economy into perspective. Environmental Innovation and Societal Transitions, 23, 3–10. doi:https://doi.org/10.1016/j.eist.2017.01.003 Frey, C., & Osborne, M. (2017). The future of employment: How susceptible are jobs to computerisation? Technological Forecasting and Social Change, 114, 254–280. doi:https://doi.org/10.1016/j.techfore.2016.08.019 Friedman, T. (2005). The world is flat: A brief history of the twenty-first century. New York, NY: Farrar, Straus & Giroux. Fuchs, C. (2014). Digital Labour and Karl Marx. London: Routledge. Funtowicz, S., & Ravetz, J. (1990). Uncertainty and quality in science for policy. Dordrecht, the Netherlands: Kluwer. Funtowicz, S., & Ravetz, J. (1991). A new scientific methodology for global environmental issues. In R. Costanza (Ed.), Ecological economics: The science and management of sustainability (pp. 137–152). New York, NY: Columbia University Press. Funtowicz, S., & Ravetz, J. (1993). Science for the post-normal age. Futures, 25(7), 739–755. doi:10.1016/0016-3287(93)90022-L Funtowicz, S., & Ravetz, J. (2008). Values and uncertainty. In G. Hirsch Hadorn, H. Hoffmann-Riem, S. Biber-Klemm, W. Grossenbacher-Mansuy, D. Joye, C. Pohl, & E. Zemp (Eds.), Handbook of transdisciplinary research (pp. 361–368). Heidelberg, Germany: Springer. Fuster Morell, M. (2011). “¿La participación como ecosistema? Comunidades de creación online parra la construcción de bienes públicos digitales.” In A. Calle Collado (Ed.), En democracia radical: Entrevínculos y utopías (pp. 129–148). Barcelona, Spain: Icaria. Galperin, H., Viecens, M., & Greppi, C. (2015). Discrimination in online contracting: Evidence from Latin America. Ottawa, Canada: International Development Research Centre and the Canadian International Development Agency. Retrieved from https:// idl-bnc.idrc.ca/dspace/bitstream/10625/55533/1/IDL-55533.pdf. Accessed on May 9, 2016. Gandini, A. (Forthcoming). Platform labour: From digital labour to labour publics. Paper Presented at “Platfrom Swindles, Platform Struggles” Seminar Series, Paris, April 20, 2018.

References    181 Gansky, L. (2010). The mesh: Why the future of business is sharing. New York, NY: Portfolio Penguin. Garben, S. (2017). Protecting workers in the online platform economy: An overview of regulatory and policy developments in the EU. Brussels, Belgium: European Agency for Safety and Health at Work. Retrieved from https://osha.europa.eu/sites/default/files/publications/documents/Protecting_Workers_in_Online_Platform_Economy.pdf. Accessed on May 2018. Garnham, N. (1979). Contribution to a political economy of mass-communication. Media, Culture & Society, 1(2), 123–146. Gash, V. (2008). Bridge or trap? Temporary workers’ transitions to unemployment and to the standard employment contract. European Sociological Review, 24(5), 651–668. doi:10.1093/esr/jcn027 Gawer, A., & Cusumano, M. (2002). Platform leadership. Boston, MA: Harvard Business School Press. Gefen, D., Benbasat, I., & Pavlou, P. (2008). A research agenda for trust in online environments. Journal of Management Information Systems, 24(4), 275–286. Gehl, R. W. (2011). The archive and the processor: The internal logic of Web 2.0. New Media & Society, 13(8), 1228–1244. Gerald, F. (2014). Workers without employers: Shadow corporations and the rise of the gig economy. Review of Keynesian Economics, 2(2), 171–188. doi:https://doi.org/10.4337/ roke.2014.02.03 Gerbaudo, P. (2014). Populism 2.0: Social media activism, the generic internet user and interactive direct democracy. In D. Trottier & C. Fuchs (Eds.), Social media, politics and the state: protests, revolutions, riots, crime and policing in the age of Facebook, Twitter and YouTube. New York, NY: Routledge. Geron, T. (2013, January 23). Airbnb and the unstoppable rise of the share economy. Forbes. Retrieved from http://www.forbes.com/sites/tomiogeron/2013/01/23/airbnband-the-unstoppable-rise-of-the-share-economy/. Accessed on May 28, 2015. Ghani, E., Kerr, W., & Stanton, C. (2014). Diasporas and outsourcing: Evidence from oDesk and India. Management Science, 60(7), 1677–1698. Giesecke, J. (2009). Socio-economic risks of atypical employment relationships: Evidence from the German labour market. European Sociological Review, 25(6), 629–646. doi:10.1093/esr/jcp012 Gillespie, T. (2010). The politics of “platforms.” New Media & Society, 12(3), 347–364. doi:10.1177/1461444809342738 Gillespie, T. (2014). The relevance of algorithms. In T. Gillespie, P. Boczkowski, & K. Foot (Eds.), Media technologies: Essays on communication, materiality, and society (inside technology) (pp. 167–194). Cambridge, MA: MIT Press. Gino, F. (2017) Uber shows how not to apply behavioural economics. Harvard Business Review. Retrieved from https://hbr.org/2017/04/uber-shows-how-not-to-apply-behavioral-economics Gobble, M. M. (2015). Regulating innovation in the new economy. Research-Technology Management, March–April, 62–63. doi:10.5437/08956308X5802005 Graham, P. (2006). Hypercapitalism: New media, language and social perceptions of value. New York, NY: Peter Lang. Graham, J., & Gandini, A. (Eds.). (2017). Collaborative production in the creative industries. London: University of Westminster Press. Graham, M., Hjorth, I., & Lehdonvirta, V. (2017). Digital labour and development: Impacts of global digital labour platforms and the gig economy on worker livelihoods. Transfer: European Review of Labour and Research, 23(2), 135–162. doi:10.1177/1024258916687250 Grassmuck, V. (2012a). The sharing turn: Why we are generally nice and have a good chance to cooperate our way out of the mess we have gotten ourselves into. In W. Sützl, F. Stalder, R. Maier, & T. Hug (Eds.), Cultures and ethics of sharing (pp. 17–34). Innsbruck, Austria: Innsbruck University Press.

182   References Grassmuck, V. (2012b). The sharing turn: Why we are generally nice and have a good chance to cooperate our way out of the mess we have gotten ourselves into. In W. Sützl, F. Stalder, R. Maier, & T. Hug (Eds.), Cultures and ethics of sharing (pp. 17–34). Innsbruck, Austria: Innsbruck University Press. Green, A., & Livanos, I. (2015). Involuntary non-standard employment in Europe. European Urban and Regional Studies, 24(2), 175–192. doi:10.1177/0969776415622257 Greenwood, B., & Wattal, S. (2015). Show me the way to go home: An empirical investigation of ride sharing and alcohol-related motor vehicle homicide. Paper No. 15-054. Fox School of Business Research. Retrieved from http://ssrn.com/abstract=2557612 Griswold, L. (2014, September 12). Uber Just caved on a big policy change after its drivers threatened to strike. Slate.com. Retrieved from http://www.slate.com/blogs/ moneybox/2014/09/12/uber_drivers_strike_they_protested_cheap_uberx_fares_uber_ backed_down.html. Accessed on January 18, 2016. Gruszka, K. (2017). Framing the collaborative economy: Voices of contestation. Environmental Innovation and Societal Transitions, 23, 92–104. Retrieved from https:// doi.org/10.1016/j.eist.2016.09.002 Guendelsberger, E. (2015, May 7). I was an undercover Uber driver. Philadelphia City Paper. Retrieved from http://citypaper.net/uberdriver/. Accessed on November 28, 2015. Gurvich, I., Lariviere, M., & Moreno, A. (2015). Operations in the On-Demand Economy: Staffing Services with Self-Scheduling Capacity. Retrieved from http://dx.doi.org/ 10.2139/ssrn.2336514 (20-08-2015) Guttentag, D. (2013). Airbnb: Disruptive innovation and the rise of an informal tourism accommodation sector. Current Issues in Tourism, 1–26. doi:10.1080/13683500.2013.827159 Guttentag, D., Smith, S., Potwarka, L., & Havitz, M. (2017). Why tourists choose Airbnb: A motivation-based segmentation study. Journal of Travel Research, 57(3), 342–359. doi:10.1177/0047287517696980 Hacker, J. (2004). Privatizing risk without privatizing the welfare state: The hidden politics of social policy retrenchment in the United States. The American Political Science Review, 98(2), 243–260. Hagiu, A. (2015). Work 3.0: Redefining Jobs And Companies In The Uber Age. Forbes. Retrieved from http://www.forbes.com/sites/hbsworkingknowledge/2015/09/29/work3-0-redefining-jobs-and-companies-in-the-uber-age/#5d00b71a7dc1. Accessed on January 28, 2016. Hagiu, A., & Biederman, R. (2015, August 21). Companies need an option between contractor and employee. Harvard Business Review. Retrieved from https://hbr.org/2015/08/ companies-need-an-option-between-contractor-and-employee. Accessed on January 28, 2016. Hagiu, A., & Wright, J. (2013). Do you really want to be an eBay? Harvard Business Review, 91(3), 102–108. Hagiu, A., & Wright, J. (2015a). Enabling versus controlling. Harvard Business School Strategy Unit Working Paper No. 16-002. Retrieved from http://ssrn.com/ abstract=2627843. Accessed on November 20, 2015. Hagiu, A., & Wright, J. (2015b). Marketplace or reseller? Management Science, 61(1), 184– 203. doi:10.1287/mnsc.2014.2042 Hagiu, A., & Wright, J. (2015c). Multi-sided platforms. International Journal of Industrial Organization. 43, 162–174. doi:10.1016/j.ijindorg.2015.03.003 Hall, P. (1993). Policy paradigms, social learning and the state: The case of economic policymaking in Britain. Comparative Politics, 25(3), 275–296. Hall, J., Horton, J., & Knoepfle, D. (2017). Labor market equilibration: Evidence from Uber. Retrieved from http://john-joseph-horton.com/papers/uber_price.pdf. Accessed on May 2018. Hall, J., & Krueger, A. (2015). An analysis of the labor market for Uber’s driver-partners in the United States. Working Papers. Princeton, NJ: Princeton University Press. Retrieved from http://dataspace.princeton.edu/jspui/bitstream/88435/dsp010z708z67d/5/587.pdf.

References    183 Hamari, J., Sjöklint, M., & Ukkonen, A. (2016). The sharing economy: Why people participate in collaborative consumption. Journal of the Association for Information Science and Technology. 67, 2047–2059. doi:10.1002/asi.23552 Hampton, K., Goulet, L., Rainie, L., & Purcell, K. (2011). Social networking sites and our lives. Washington, DC: Pew Research Center. Retrieved from http://www.pewinternet. org/files/old-media/Files/Reports/2011/PIP - Social networking sites and our lives.pdf. Accessed on August 19, 2015. Hancké, B., Rhodes, M., & Thatcher, M. (2008). Beyond varieties of capitalism. In B. Hancké (Ed.), Debating varieties of capitalism. Oxford: Oxford University Press. Hantke-Domas, M. (2003). The public interest theory of regulation: Non-existence or misinterpretation? European Journal of Law and Economics, 15(2), 165–194. doi:10.1023/A:1021814416688 Harman, C. (2010) Zombie capitalism: Global crisis and the relevance of Marx. Chicago, IL: Haymarket. Harris, S., & Krueger, A. (2015). A proposal for modernizing labor laws for twenty-firstcentury work: The “independent worker.” Discussion Paper 2015-10. Washington, DC: The Hamilton Project, Brookings. Retrieved from http://www.hamiltonproject.org/ assets/files/modernizing_labor_laws_for_twenty_first_century_work_krueger_harris. pdf. Accessed on January 11, 2016. Hausemer, P., ìRzepecka, J., Dragulin, M., Vitiello, S., Rabuel, L., Nunu, M., Diaz, A., Psaila, E., Fiorentini, S., Gysen, S., Meeusen, T., Quaschning, S., Dunne, A., Grinevich, V., Huber, F., & Baines, L. (2017). Exploratory study of consumer issues in online peer-topeer platform markets. Final Report. Brussels, Belgium: Directorate-General for Justice and Consumers, European Commission. Head, B. (2013). Evidence-based policymaking: Speaking truth to power? Australian Journal of Public Administration, 72(4), 397–403. doi:10.1111/1467-8500.12037 Heimans, J., & Timms, H. (2014). Understanding “new power.” Harvard Business Review, 93(12), 48–56. Heinrichs, H. (2013). Sharing economy: A potential new pathway to sustainability. GAIA, 22(4), 228–231. Herbig, B., Dragano, N., & Angerer, P. (2013). Health in the long-term unemployed. Dtsch Arztebl Int, 110(23–24), 413–419. Hesmondhalgh, D. (2010). User-generated content, free labour and the cultural industries. Ephemera, 10(3/4), 267–284. Hill, S. (2015a). New economy, new social contract: A plan for a safey net in a multiemployer world. Washington, DC: New America Foundation. Retrived from https://static. newamerica.org/attachments/4395-new-economy-new-social-contract/New Economy, Social Contract_UpdatedFinal.34c973248e6946d0af17116fbd6bb79e.pdf. Accessed on November 20, 2015. Hill, S. (2015b). Raw deal: How the “Uber economy” and runaway capitalism are screwing American workers. New York, NY: St. Martin’s Press. Hipp, L., Bernhardt, J., & Allmendinger, J. (2015). Institutions and the prevalence of nonstandard employment. Socio-Economic Review, 13(2), 351–377. doi:10.1093/ser/mwv002 Hirschman, A. O. (1977). The passions and the interests: Political arguments for capitalism before its triumph. Princeton, NJ: Princeton University Press. Hirschman, A. O. (1982). Shifting involvements: Private interest and public action. Princeton, NJ: Princeton University Press. Hirschman, A. O. (1991). The Rhetoric of Reaction. Cambridge, MA: Harvard University Press. Hogg, T., & Adamic, L. (2004). Enhancing reputation mechanisms via online social networks. EC ‘04 proceedings of the 5th ACM conference on electronic commerce (Vol. 5, pp. 236–237). New York, NY: ACM. Holtgrewe, U. (2014). New new technologies: The future and the present of work in information and communication technology. New Technology, Work and Employment, 29(1), 9–24. doi:10.1111/ntwe.12025

184   References Hong, Y., & Pavlou, P. (2014). Is the world truly “flat”? Empirical evidence from online labor markets. Fox School of Business Research Paper No. 15-045. Retrieved from http://ssrn. com/abstract=2371748. Accessed on December 2, 2015. Hong, Y., Wang, C., & Pavlou, P. (2014). Comparing open and sealed bid auctions: Evidence from online labor markets. NET Institute Working Paper No. 13-05, Information Systems Research. Retrieved from http://ssrn.com/abstract=2335669. Accessed on December 2, 2015. Horton, J. (2010). Online labor markets. Internet and Network Economics, 6484, 515–522. doi:10.1007/978-3-642-17572-5_45 Horton, J. (2011). The condition of the Turking class: Are online employers fair and honest? Economics Letters, 111(1), 10–12. Retrieved from http://dx.doi.org/10.1016/j.econlet.2010.12.007 Horton, J. (2014). Misdirected search effort in a matching market: Causes, consequences and a partial solution. Paper presented at the Proceedings of the 15th ACM conference on economics and computation, Palo Alto, CA. Horton, J. (2015a). The effects of algorithmic labor market recommendations: Evidence from a field experiment. Retrieved from http://ssrn.com/abstract=2346486. Accessed on November 28, 2015. Horton, J. (2015b). Supply constraints as a market friction: Evidence from an online labor market. Retrieved from http://john-joseph-horton.com/papers/scs.pdf. Accessed on November 28, 2015. Horton, J., & Chilton, L. (2010). The labor economics of paid crowdsourcing. Paper presented at the Proceedings of the 11th ACM conference on electronic commerce, Cambridge, MA. Horton, J., & Golden, J. (2015). Reputation inflation: Evidence from an online labor market. Retrieved from http://econweb.tamu.edu/common/files/workshops/Theory and Experimental Economics/2015_3_5_John_Horton.pdf Horton, J., & Johari, R. (2015). At what quality and what price?: Eliciting buyer preferences as a market design problem. Paper presented at the Proceedings of the 16th ACM conference on economics and computation, Portland, OR. Horton, J., & Tambe, P. (2015). Labor economists get their microscope: Big data and labor market analysis. Big Data, 3(3), 130–137. doi:10.1089/big.2015.0017 Horton, J., & Zeckhauser, R. (2016a). Owning, using and renting: Some simple economics of the “sharing economy.” (p. 19525). NBER Working Paper Series No. 22029. Retrieved from http://www.nber.org/papers/w22029. Accessed on June 2018. Horton, J., & Zeckhauser, R. (2016b). Owning, using, and renting: Some simple economics of the “sharing economy.” NBER Working Paper Series, 22029. Retrieved from http:// www.nber.org/papers/w22029 House of Lords. (2016). Online platforms and the digital single market. Oral and written evidence. London: House of Lords, Select Committee on the European Union Internal Market Sub-Committee. Retrieved from https://www.parliament.uk/ documents/lords-committees/eu-internal-market-subcommittee/online-platforms/ OnlinePlatformsWrittenEvVolumePublished.pdf. Accessed on April 2018. Howcroft, D., & Taylor, P. (2014). “Plus ca change, plus la meme chose?” – Researching and theorising the “new” new technologies. New Technology, Work and Employment, 29(1), 1–8. doi:10.1111/ntwe.12026 Howe, J. (2006, June 1). The rise of crowdsourcing. Wired. Retrieved from http://www.wired. com/2006/06/crowds/. Accessed on December 1, 2015. Howe, J. (2008). Crowdsourcing – How the power of the crowd is driving the future of business. New York, NY: Crown Publishing Group. Huws, U., & Joyce, S. (2016a). Size of the Sweden’s “gig economy” revealed for the first time University of Hertfordshire and UNI Europa. Retrieved from http://www.feps-europe.eu/ assets/3f853cec-1358-4fb4-9552-274b55e05ecf/crowd-working-survey-swedenpdf.pdf. Accessed on April 19, 2016.

References    185 Huws, U., & Joyce, S. (2016b). Size of the UK’s “gig economy” revealed for the first time University of Hertfordshire and UNI Europa. Retrieved from http://www.feps-europe. eu/assets/a82bcd12-fb97-43a6-9346-24242695a183/crowd-working-surveypdf.pdf. Accessed on April 19, 2016. Huws, U., Spencer, N., Syrdal, D., & Holts, K. (2017). Work in the European gig economy: Research results from the UK, Sweden, Germany, Austria, the Netherlands, Switzerland and Italy. Hertfordshire, UK: FEPS, UniGlobal and University of Hertfordshire. Ichino, A., Mealli, F., & Nannicini, T. (2008). From temporary help jobs to permanent employment: what can we learn from matching estimators and their sensitivity? Journal of Applied Econometrics, 23(3), 305–327. doi:10.1002/jae.998 International Labour Office (ILO). (1997). Part-time work: Solution or trap? International Labour Review, 136(4), 557–579. International Labour Office (ILO). (2015). Non-standard forms of employment. Geneva, Switzerland: ILO. International Labour Office (ILO). (2016). Non-standard employment around the world: Understanding challenges, shaping prospects. Geneva, Switzerland: ILO. International Labour Office (ILO). (2017). World social protection report 2017–19: Universal social protection to achieve the sustainable development goals. Geneva, Switzerland: ILO. Ipeirotis, P. (2010). Demographics of mechanical Turk. New York, NY: Department of Information, Operation, and Management Sciences, New York University. Retrieved from https://archive.nyu.edu/bitstream/2451/29585/2/CeDER-10-01.pdf. Accessed on December 4, 2015. Ipeirotis, P., & Horton, J. (2011). The need for standardization in crowdsourcing. Paper presented at the CHI 2011, Vancouver, BC, Canada. Irani, L. (2015). The cultural work of microwork. New Media & Society, 17(5), 720–739. Irani, L., & Silberman, M. (2013). Turkopticon: interrupting worker invisibility in Amazon mechanical turk. Paper presented at the Proceedings of the SIGCHI conference on human factors in computing systems, Paris, France. Jacobs, A. (2011). Governing for the long term: Democracy and the politics of investment. Cambridge: Cambridge University Press. Janine, L. (2009). The segmentation potential of non-standard employment: A fourcountry comparison of mobility patterns. International Journal of Manpower, 30(7), 692–715. doi:10.1108/01437720910997353 Jen, R., Kaur, J., De Heus, J., & Dillahunt, T. (2014). Analyzing employment technologies for economically distressed individuals. Paper presented at the CHI ‘14 extended abstracts on human factors in computing systems, Toronto, Ontario, Canada. Jenkins, H. (2006). Convergence culture: Where old and new media collide. New York, NY: New York University Press. John, M. K. (1963[1930]). Essays in Persuasion (pp. 358-373), New York: W.W. Norton & Co. John, N. A. (2013a). Sharing and Web 2.0: The emergence of a keyword. New Media and Society, 15(2), 167–182. doi:10.1177/1461444812450684 John, N. A. (2013b). The social logics of sharing. Communication Review, 16(3), 113–131. doi:10.1080/10714421.2013.807119 Jones, B., & Baumgartner, F. (2005). The politics of attention: How government prioritizes problems. Chicago, IL: University of Chicago Press. Jones, B. D., Thomas, H. F., & Wolfe, M. (2014). Policy bubbles. Policy Studies Journal, 42(1), 146–171. doi:10.1111/psj.12046 Jøsang, A., Ismail, R., & Boyd, C. (2007). A survey of trust and reputation systems for online service provision. Decision Support Systems, 43(2), 618–644. doi:10.1016/j. dss.2005.05.019 Kahn, L. M. (2012). Labor market policy: A comparative view on the costs and benefits of labor market flexibility. Journal of Policy Analysis and Management, 31(1), 94–110. doi:10.1002/pam.20602

186   References Kalleberg, A. L. (2009). Precarious work, insecure workers: Employment relations in transition. American Sociological Review, 74(1), 1–22. doi:10.1177/000312240907400101 Kang, C., & Eilperin, J. (2015, 28 February). Why Silicon Valley is the new revolving door for Obama staffers. The Washington Post. Retrieved from http://www.washingtonpost.com/business/economy/as-obama-nears-close-of-his-tenure-commitment-tosilicon-valley-is-clear/2015/02/27/3bee8088-bc8e-11e4-bdfa-b8e8f594e6ee_story.html. Accessed on May 28, 2015. Kantor, J. (2014, August 13). Working anything but 9 to 5. The New York Times. Retrieved from http://www.nytimes.com/interactive/2014/08/13/us/starbucks-workers-schedulinghours.html?_r=0. Accessed on November 30, 2015. Karatzogianni, (2006). The Politics of Cyberconflict, London and New York, NY: Routledge. Karatzogianni, A. (2015). Firebrand waves of digital activism 1994–2014: The rise and spread of hacktivism and cyberconflict. Basingstoke: Palgrave Macmillan. Kässi, O., & Lehdonvirta, V. (2016). Online labour index: Measuring the online gig economy for policy and research. Paper presented at Internet, Politics & Policy 2016, Oxford, 22–23 September. Retrieved from http://ilabour.oii.ox.ac.uk/online-labour-index/ Kaufmann, N., Schulze, T., & Veit, D. (2011). More than fun and money. Worker motivation in crowdsourcing: A study on mechanical turk. All Submissions, Paper 340. In AMCIS 2011 proceedings. Retrieved from http://aisel.aisnet.org/amcis2011_submissions/340. Accessed on December 1, 2015. Kautonen, T., Down, S., Welter, F., Vainio, P., Palmroos, J., Althoff, K., & Kolb, S. (2010). “Involuntary self-employment” as a public policy issue: A cross-country European review. International Journal of Entrepreneurial Behavior & Research, 16(2), 112–129. doi:10.1108/13552551011027002 Kay, A. (2011). Evidence-based policy-making: The elusive search for rational public administration. Australian Journal of Public Administration, 70(3), 236–245. doi:10.1111/j.1467-8500.2011.00728.x Keating, D. (2014, September 11). An Uber­sensitive subject. European Voice. Retrieved from http://www.politico.eu/other-voices/an-uber-sensitive-subject/. Accessed on May 31, 2015. Kennedy, J. (2016). Conceptual boundaries of sharing. Information, Communication & Society, 19(4), 461–474. doi:10.1080/1369118X.2015.1046894 Kenney, M., & Zysman, J. (2016). What is the future of work? Understanding the platform economy and computation-intensive automation. In Berkeley roundtable on the international economy, conference on work and welfare in the platform economy, Proto-paper prepared for Radcliffe Institute. Kerr, D. (2014, August 20). Vexed in the city: The “sharing” economy’s hidden toll on San Francisco. CNET. Retrieved from http://www.cnet.com/news/vexed-in-the-city-thesharing-economys-hidden-toll-on-san-francisco/. Accessed on May 30, 2015. Kessler, S. (2015, February 17). The gig economy won’t last because it’s being sued to death. Fast Company. Retrieved from http://www.fastcompany.com/3042248/the-gigeconomy-wont-last-because-its-being-sued-to-death Accessed on January 21, 2016. Keynes, J. M. (1963 [1930]). Essays in persuasion (pp. 358–373). New York, NY: W. W. Norton. King, S. P. (2015). Sharing economy: What challenges for competition law? Journal of European Competition Law & Practice, 6(10), 729–734. doi:10.1093/jeclap/lpv072 Klandermans, B. (1997). The social psychology of protest. Oxford: Blackwell. Kleemann, K., Voß, G., & Rieder, K. (2008). Un(der)paid innovators: The commercial utilization of consumer work through crowdsourcing. Science, Technology and Innovation Studies, 4(1), 5–26. Koopman, C., Mitchell, M., & Thierer, A. (2014). The sharing economy and consumer protection regulation: The case for policy change. Arlington, TX: Mercatus Center, George Madison University. Retrieved from http://mercatus.org/sites/default/files/KoopmanSharing-Economy.pdf. Accessed on May 28, 2015.

References    187 Koopman, C., Mitchell, M., & Thierer, A. (2015). The sharing economy: Issues facing platforms, participants, and regulators. Sharing Economy Workshop, Project No. P15-1200. Fairfax, VA: Mercatus Center at George Mason University. Koso, M. (2015. Companies Like Uber Could Learn a Thing or Two from This Office Cleaning Startup, Where the Workers Are as Happy as the Clients. Business Insider. Retrieved from http://www.businessinsider.com/managed-by-q-hires-cleaners-as-employees2015-3#ixzz3dXpUNKru. Accessed on January 21, 2016. Kroet, C. (2014, December 11). Uber’s ride­sharing service faces ban in Spain and the Netherlands. European Voice. Retrieved from http://www.politico.eu/article/dutchcourt-bans-ubers-ride-sharing-service/. Accessed on May 31, 2015. Kuek, S.-C. , Paradi-Guilford, C., Fayomi, T., Imaizumi, S., & Ipeirotis, P. (2015). The global opportunity in online outsourcing. Washington DC: World Bank Group. Kuttner, R. (2013). Thetask rabbit economy. The American Prospect, September–October, pp. 46–55. Kuttner, R. (2016). The new inequality debate. American Prospect, Winter Available at https://pages.wustl.edu/files/pages/imce/fazz/kuttner_new_inequality.pdf. Kwan, M., & Ramachandran, D. (2009). Trust and online reputation systems. In J. Golbeck (Ed.), Computing with social trust (pp. 287–312). London: Springer. Lamberton, C., & Rose, R. (2012). When is ours better than mine? A framework for understanding and altering participation in commercial sharing systems. Journal of Marketing, 76(4), 109–125. doi:10.1509/jm.10.0368 Landier, A., Szomoru, D., & Thesmar, D. (2016). Working in the on demand economy: An analysis of Uber driver partners in France. Retrieved from http://mitsloan.mit.edu/ shared/ods/documents/?DocumentID=2529. Accessed on May 2018. Lang, N. (2015, November 22). The real reason on-demand startups are reclassifying workers. Retrieved from http://techcrunch.com/2015/11/22/the-real-reason-on-demandstartups-are-reclassifying-workers/. Accessed on November 29, 2015. Laurens, S. (2017). Lobbyists and bureaucrats in Brussels: Capitalism’s brokers. London: Routledge. Lauterbach, D., Truong, H., Shah, T., & Adamic, L. (2009). Surfing a Web of trust: Reputation and reciprocity on Couchsurfing.com. Paper presented at the International conference on computational science and engineering, August 29–31. Le Crosnier, H. (2015). En communs: Une introduction aux communs de la connaissance. Caen, C&F, coll. Blogollection. Leadbeater, C. (2009). We think: Mass innovation, not mass production. London: Profile Books. Lee, C. (2015). The sharers’ gently-used clothes. Contexts, 14(1), 17–18. Lee, M. K., Kusbit, D., Metsky, E., & Dabbish, L. (2015). Working with machines: The impact of algorithmic and data-driven management on human workers. Paper presented at the Proceedings of the 33rd annual ACM conference on human factors in computing systems, Seoul, Republic of Korea. Lehdonvirta, V. (2016). Algorithms that divide and unite: Delocalization, identity, and collective action in “microwork.” In J. Flecker (Ed.), Space, place and global digital work. London: Palgrave-Macmillan. Retrieved from http://vili.lehdonvirta.com/files/ Lehdonvirta%202016%20Delocalization%20identity%20collective%20action%20 in%20microwork.pdf Lehdonvirta, V., Barnard, H., Graham, M., & Hjorth, I. (2014). Online labour markets – Leveling the playing field for international service markets? Paper presented at the IPP2014: Crowdsourcing for politics and policy, Oxford. Retrieved from: http://ipp.oii.ox.ac.uk/ sites/ipp/files/documents/IPP2014_Lehdonvirta_0.pdf. Accessed on October 15, 2015. Lessig, L. (1999). “Code is law,” code and other laws of cyberspace. London: Basic Books. Retrieved from http://pdf.codev2.cc/Lessig-Codev2.pdf Li, J. (2015). Is online media a two-sided market? Computer Law & Security Review: The International Journal of Technology Law and Practice, 31(1), 99–111. doi:10.1016/ j.clsr.2014.11.001

188   References Liber, R. (2013, May 21). A $2,400 fine for an Airbnb host. The New York Times. Retrieved from http://bucks.blogs.nytimes.com/2013/05/21/a-2400-fine-for-an-airbnb-host/?_r=0. Accessed on May 30, 2015. Lien, T. (2015). Luxe Valet Turning its Independent Contractors into Employees. Los Angeles Times. Retrieved from http://www.latimes.com/business/technology/la-fi-tnluxe-employee-20150730-story.html. Accessed on January 21, 2016. Loader, B. (Ed.). (1998). Cyberspace divide: Equality, agency and policy in the information society. London: Routledge. Lougher, G., & Kalmanowicz, S. (2016). EU competition law in the sharing economy. Journal of European Competition Law & Practice, 7(2), 87–102. doi:10.1093/jeclap/lpv086 Luchetta, G. (2014). Is the Google platform a two-sided market? Journal of Competition Law and Economics, 10(1), 185–207. doi:10.1093/joclec/nht026 Madden, S. (2015, July 31). Why homejoy failed … and the future of the on-demand economy. Retrieved from http://techcrunch.com/2015/07/31/why-homejoy-failed-and-thefuture-of-the-on-demand-economy/. Accessed on December 1, 2015. MacLeod, W., & Malcomson, J. (1998). Motivation and markets. The American Economic Review, 88(3), 388–411. Malhotra, A., & Van Alstyne, M. (2014). The dark side of the sharing economy … and how to lighten it. Communications of the ACM, 57, 24–27. doi:10.1145/2668893 Malmendier, U., & Schmidt, K. (2012). You owe me. NBER Working Paper Series, No. 18543. doi:10.3386/w18543 Maor, M. (2014). Policy persistence, risk estimation and policy underreaction. Policy Sciences, 47(4), 425–443. Maor, M. (2016). Emotion-driven negative policy bubbles. Policy Sciences, 49(2), 191–210. doi:10.1007/s11077-015-9228-7 Martin, C. J. (2016). The sharing economy: A pathway to sustainability or a nightmarish form of neoliberal capitalism? Ecological Economics, 121, 149–159. doi:10.1016/ j.ecolecon.2015.11.027 Martin, D., Hanrahan, B., O’Neill, J., & Gupta, N. (2014). Being a turker. Paper presented at the Proceedings of the 17th ACM conference on computer-supported cooperative work & social computing, Baltimore, MD. Martin, E., & Shaheen, S. (2010). Greenhouse gas impacts of car sharing in North America. Report 09-11. San Jose, CA: Mineta Transportation Institute. Martin, C. J., & Upham, P. (2016). Grassroots social innovation and the mobilisation of values in collaborative consumption: A conceptual model. Journal of Cleaner Production. 134(Part A), 204–213. Retrieved from http://dx.doi.org/10.1016/j.jclepro.2015.04.062 Marvit, M. (2014, February 5). How crowdworkers became the ghosts in the digital machine. The Nation. Retrieved from http://www.thenation.com/article/how-crowdworkers-became-ghosts-digital-machine/. Accessed on January 21, 2016. Marx, C. (1959). Capital III. Moscow, Russia: Foreign Languages Publishing House. Mason, W., & Watts, D. J. (2010). Financial incentives and the “performance of crowds.” SIGKDD Explorations Newsletter, 11(2), 100–108. doi:10.1145/1809400.1809422 Masseli, I., Lenaerts, K., & Beblavý, M. (2016). Five things we need to know about the on-demand economy. Brussels, Belgium: Centre for European Policy Studies. Retrieved from https:// www.ceps.eu/system/files/CEPS%20Essay%20No%2021%20On%20Demand%20 Economy.pdf Matsaganis, M., Özdemir, E., Ward, T., & Zavakou, A. (2016). Non-standard employment and access to social security benefits. Research Note 8/2015. Brussels, Belgium: European Commission, Directorate General for Employment, Social Affairs and Inclusion. Matthews, J. (2017). Beyond “collaborative economy” discourse: Present, past and potential of intermediation platforms. In J. Graham & A. Gandini, A. (Eds.), Collaborative production in the creative industries. London: University of Westminster Press. Matthews, J., & Smith Maguire, J. (2014). The cultural intermediaries reader. London: SAGE.

References    189 Matzler, K., & Kathan, W. (2015). Adapting to the sharing economy. MIT Sloan Management Review, 56(2), 71–77. Mayzlin, D., Dover, Y., & Chevalier, J. (2014). Promotional reviews: An empirical investigation of online review manipulation. American Economic Review, 104(8), 2421–2455. doi:10.1257/aer.104.8.2421. MBO Partners. (2015). Independent workers and the on-demand economy. Herndon, VA: MBO Partners. Retrieved from http://info.mbopartners.com/rs/mbo/images/ On-Demand-Economy-2014.pdf. Accessed on June 24, 2015. McCaughey, M., & Ayers, M. (2003). Cyberactivism: Online activism in theory and practice. London: Routledge. McKinsey. (2016). Independent work: Choice, necessity, and the gig economy. Brussels, Belgium: McKinsey Global Institute. McLean, S. (2015). The rise of the sharing economy. Computers & Law Magazine of SCL, 26(1), 26–28. McNeill, D. (2016). Governing a city of unicorns: Technology capital and the urban politics of San Francisco. Urban Geography, 37(4), 494–513. doi:10.1080/02723638.2016.1139868 Meelen, T., & Frenken, K. (2015, January 14). Stop saying Uber is part of the sharing economy. FastCompany. Retrieved from http://www.fastcoexist.com/3040863/stop-saying-uber-is-part-of-the-sharing-economy. Accessed on May 28, 2015. Milan, S. (2013). Social movements and their technologies: Wiring social change. New York, NY: Palgrave MacMillan. Mill, R. (2011). Hiring and learning in online global labor markets. NET Institute Working Paper No. 11-17. Available at SSRN: https://ssrn.com/abstract=1957962 or http:// dx.doi.org/10.2139/ssrn.1957962. Miller, N., Resnick, P., & Zeckhauser, R. (2005). Eliciting informative feedback: The peerprediction method. Management Science, 51(9), 1359–1373. doi:10.2307/20110426 Miller, S. (2014). Transferable sharing rights: A theoretical model for regulating airbnb and the short-term rental market. Retrieved from http://ssrn.com/abstract=2514178. Accessed on May 20, 2015. Miller, S. (2015). First principles for regulating the sharing economy. Harvard Journal on Legislation. 53, 147–202. Retrieved from http://ssrn.com/abstract=2568016. Accessed on May 23, 2015. Miller, S. (2016). First principles for regulating the sharing economy. Harvard Journal on Legislation, 53, 147–202. Minter, K. (2017). Negotiating labour standards in the gig economy: Airtasker and Unions New South Wales. The Economic and Labour Relations Review, 28(3), 438–454. doi:10.1177/1035304617724305 Misuraca, G., Codagnone, C., & Rossel, P. (2013). From practice to theory and back to practice: Reflexivity in measurement and evaluation for evidence-based policy making in the information society. Government Information Quarterly, 30(1), S68–S82. doi:10.1016/j.giq.2012.07.011 Möhlmann, M. (2015). Collaborative consumption: Determinants of satisfaction and the likelihood of using a sharing economy option again. Journal of Consumer Behaviour, 14(3), 193–207. doi:10.1002/cb.1512 Moore, P. (2017). The quantified self in precarity: work, technology and what counts. Abingdon: Routledge. Morgan, B., & Kuch, D. (2015). Radical transactionalism: Legal consciousness, diverse economies, and the sharing economy. Journal of Law and Society, 42(4), 556–587. doi:10.1111/j.1467-6478.2015.00725.x Murray, C. (1984). Losing Ground: American Social Policy, 1950–1980. New York Basic Books. Musthag, M., & Ganesan, D. (2013). Labor dynamics in a mobile micro-task market. Paper presented at the Proceedings of the SIGCHI conference on human factors in computing systems, Paris, France. Nadeem, S. (2015). On the sharing economy. Contexts, 14(1), 13.

190   References Neeser, D. (2015). Does Airbnb hurt hotel business: Evidence from the Nordic countries. Máster Universitario en Análisis Económico, Universidad Carlos III de Madrid, Madrid, Spain. Retrieved from https://www.researchgate.net/publication/282151529. Accessed on March 20, 2016. Newton, C. (2014, June 17). “TaskRabbit is blowing up its business model and becoming the Uber for everything.” The Verge. Retrieved from http://www.theverge. com/2014/6/17/5816254/taskrabbit-blows-up-its-auction-house-to-offer-services-on-demand. Accessed on November 30, 2015. Nielsen, P. (2015). “Sharing economy” in Denmark – First attempt to measure. Statistics Denmark. Retrieved from http://dynamicsofvirtualwork.com/wp-content/ uploads/2016/03/Nielsen-presentation.pdf Nixon, B. (2014). Toward a political economy of “audience labour” in the digital era. Triple C, 12(2), 713–734. Nixon, B. (2017). Critical communication policy research and the attention economy: From digital labor theory to digital class struggle. International Journal of Communication, 11, Feature 1–12. Nosko, C., & Tadelis, S. (2015). The limits of reputation in platform markets: An empirical analysis and field experiment. NBER Working Paper Series, 20830. doi:10.3386/w20830 Nubelo. (2014). Primer informe regional sobre la industria del Trabajo 3.0: Nubelo. Retrieved from http://www.idigital.cat/documents/10501/25360/Estudi-Nubelo. Accessed on April 25, 2016. Nunez, I., & Livanos, I. (2014). Temps “by choice?” An investigation of the reasons behind temporary employment among young workers in Europe. Journal of Labor Research, 36(1), 44–66. doi:10.1007/s12122-014-9195-3 Nutley, S., Morton, S., Jung, T., & Boaz, A. (2010). Evidence and policy in six European countries: Diverse approaches and common challenges. Evidence & Policy, 6(2), 131–144. Observatorio Cetelem. (2013). Consumidores europeos en modo alternativo. Madrid, Spain: Observatorio Cetelem. Retrieved from http://www.elobservatoriocetelem.es/observatorio/Observatorio-consumo-europeo-2013.pdf. Accessed on August 17, 2015. O’Donovan, C. (2015). What Happens When a Delivery Service Tries to Pay its Workers Well?, TechCrunch. Retrieved from http://www.buzzfeed.com/carolineodonovan/ can-mun#.soXB8oA8W. Accessed on January 21, 2016. OECD. (2010). OECD employment outlook 2010: Moving beyond the jobs crisis. Paris, France: OECD. OECD. (2014). OECD employment outlook 2014. Paris, France: OECD Publishing. OECD. (2015a). In it together: Why less inequality benefits all. Paris, France: OECD Publishing. OECD. (2015b). OECD employment outlook 2015. Paris, France: OECD Publishing. OECD. (2016). Working party on measurement and analysis of the digital economy. New forms of work in the digital economy. Paris, France: OECD. Oei, S., & Ring, D. (2015). Can sharing be taxed? Boston College Law School Legal Studies Research Paper No. 352. Boston, MA. Retrieved from http://ssrn.com/abstract=2570584 Office for National Statistics (ONS). (2016). The feasibility of measuring the sharing economy. London: ONS. Retrieved from https://www.ons.gov.uk/economy/economicoutputandproductivity/output/articles/thefeasibilityofmeasuringthesharingeconomy/2016-04-05/ pdf. Accessed on April 19, 2016. Olson, M. (1971 [1965]). The logic of collective action. Public goods and the theory of groups. Cambridge, MA: Harvard University Press. Ong, J., & Cabañes, J. (2017). Architects of networked disinformation: Behind the scenes of troll accounts and fake news production in the Philippines. Retrieved from http:// newtontechfordev.com/wp-content/uploads/2018/02/architects-of-networked-disinformation-full-report.pdf O’Regan, M. (2009). New technologies of the self and social networking sites: Hospitality exchange clubs and the changing nature of tourism and identity. In Y. Abbas &

References    191 F. Dervin (Eds.), Digital technologies of the self, newcastle upon Tyme (pp. 171–198). Cambridge: Cambridge Scholars. Overgoor, J., Wulczyn, E., & Potts, C. (2012). Trust propagation with mixed-effects models. Proceedings of the sixth international AAAI conference on weblogs and social media. Retrieved from http://www.aaai.org/ocs/index.php/ICWSM/ICWSM12/paper/ viewFile/4627Jan/5051 Owyang, J. (2013). The collaborative economy. An Altimeter Group Research Report (pp. 1–27), New York, NY: Altimeter Research. Owyang, J. (2015, April 23). How the collaborative economy is impacting the energy sector. Retrieved from http://www.web-strategist.com/blog/category/sharing-economy/. Accessed on June 5, 2015. Owyang, J., Samuel, A., & Grenville, A. (2014). Sharing is the new buying: Vision critical & crowd companies. Retrieved from http://tms.visioncritical.com/sites/default/files/pdf/ sharing-new-buying-collaborative-economy-report.pdf. Accessed on August 15, 2015. Ozanne, L., & Ballantine, P. (2010). Sharing as a form of anti-consumption? An examination of toy library users. Journal of Consumer Behaviour, 9(6), 485–498. doi:10.1002/cb.334 Ozanne, J., & Ozanne, L. (2011). Innovative community exchange systems: Grassroots social experiments in sustainability. Advances in Consumer Research, 39, 65–69. Pais, I., Gandini, A., & Arcidiacono, D. L. (2018). Unboxing the sharing economy: Opportunities and risks of the era of collaboration. Thousand Oaks, CA: Sociological Review Press. Pallais, A. (2013). Inefficient hiring in entry-level labor markets. NBER Working Paper Series, No. 18917. doi:10.3386/w18917 Pallais, A., & Sands, E. (2016). Why the referential treatment? Evidence from field experiments on referrals. Journal of Political Economy, 124(6), 1793–1828. doi:10.1086/688850 Parigi, P., & State, B. (2014). Disenchanting the world: The impact of technology on relationships. In L. Aiello & D. McFarland (Eds.), Social Informatics (Vol. 8851, pp. 166–182). New York, NY: Springer International Publishing. Parigi, P., State, B., Dakhlallah, D., Cook, K., & Corten, R. (2013). A community of strangers: The dis-embedding of social ties. PLoS ONE, 8(7). doi:10.1371/journal.pone.0067388 Parker, G., & Alstyne, M. W. V. (2000). Internetwork externalities and free information goods. Proceedings of the 2nd ACM conference on electronic commerce, Minneapolis, MN. Parker, G. G., & Alstyne, M. W. V. (2005). Two-sided network effects: A theory of information product design. Management Science, 51(10), 1494–1504. doi:10.2307/20110438 Paul, K., & Moser, K. (2009). Unemployment impairs mental health: Meta-analyses. Journal of Vocational Behavior, 74, 264–282. Pavlou, P., & Gefen, D. (2004). Building effective online marketplaces with institutionbased trust. Information Systems Research, 15(1), 37–59. Pawson, R., & Tilley, N. (1997). Realistic evaluation. London: SAGE Publications Ltd. Peers, S. (2013). Equal treatment of atypical workers: A new frontier for EU law? Yearbook of European Law, 32(1), 30–56. doi:10.1093/yel/yet002 Pesole, A., Urzí Brancati, M. C, Fernández-Macías, E., Biagi, F., & González Vázquez, I. (2018). Platfrom workers in Europe. Evidence from the COLLEEM Survey. Joint Research Centre Luxembourg: Publications Office of the European Union. Peters, M. A., & Bulut, E. (Eds.). (2011). Cognitive capitalism, education, and digital labor. New York, NY: Peter Lang. Petrongolo, B., & Pissarides, C. A. (2001). Looking into the black box: A survey of the matching function. Journal of Economic Literature, 39(2), 390–431. Petrongolo, B., & Pissarides, C. (2006). Scale effects in markets with search. The Economic Journal, 116(508), 21–44. doi:10.1111/j.1468-0297.2006.01046.x Pew. (2014). Millennials in adulthood. Washington, DC: Pew Research Center. Retrieved from http://www.pewsocialtrends.org/files/2014/03/2014-03-07_generations-reportversion-for-web.pdf. Accessed on August 19, 2015.

192   References Pielke, R. (2007). The honest broker: Making sense of science in policy and politics. Cambridge: Cambridge University Press. Pigou, A. (1920). The economics of welfare. New York, NY: Palgrave Macmillan. Pigou, A. (1932). The economics of welfare. London: Macmillan & Co. Pilz, D., & Gewald, H. (2013). Does money matter? Motivational factors for participation in paid- and non-profit-crowdsourcing communities. Paper presented at the 11th International Conference on Wirtschaftsinformatik, Leipzig, Germany. Retrieved from http://www. wi2013.de/proceedings/WI2013 – Track 4 - Pilz.pdf. Accessed on December 3, 2015. PIPAME. (2015). Enjeux et perspectives de la consommation collaborative. Rapport Final, République Français, Paris, France. Ministère Ministère de l’Economie, de l’Industrie, et du Numérique, Pôle Interministériel de Prospective et d’Anticipation des Mutations Economiques (PIPAME). Retrieved from http://www.entreprises.gouv.fr/files/files/directions_services/etudes-et-statistiques/prospective/Numerique/2015-07-Consommationcollaborative-Rapport-final.pdf. Accessed on February 11, 2016. Posen, H. (2016). Ridesharing in the sharing economy: Should regulators impose Über regulations on Uber? Iowa Law Review, 101, 404–433. Posner, R. (1974). Theories of economic regulation. The Bell Journal of Economics and Management Science, 5(2), 335–358. doi:10.2307/3003113 Prassl, J., & Risak, M. (2016). Uber, Taskrabbit, & Co: Platforms as employers? Rethinking the legal analysis of crowdwork. Comparative Labor Law & Policy Journal. Retrieved from https://ssrn.com/abstract=2733003. Psaila, E., Fiorentini, S., Santos Silva, M., & Gomez, A. (2017). Exploratory study of consumer issues in online peer-to-peer platform markets. Task 5 – Legal Analysis Report. Brussels, Belgium: European Commission. Retrieved from http://ec.europa.eu/newsroom/document.cfm?doc_id=45244. Accessed on May 2018. Putnam, R. (1993). Marking democracy work: Civic traditions in modern Italy. Princeton, NJ: Princeton University Press. Putnam, R. (2000). Bowling alone: The collapse and revival of American community. New York, NY: Simon & Schuster. Quattrone, G., Proserpio, D., Quercia, D., Capra, L., et al. (2016). Who Benefits from the “Sharing” Economy of Airbnb? Paper presented at the Proceedings of the 25th International Conference on World Wide Web, Montreal, Quebec, Canada. Quittner, J. (2015). How peer-to-peer worker services are changing the debate on wages. Retrieved from http://www.inc.com/jeremy-quittner/freelance-sites-add-worker-protections-like-higher-minimum-wage.html. Accessed on October 20, 2015. Raja, S., Imaizumi, S., Kelly, T., & Paradi-Guilford, C. (2013). Connecting to work. How information and communication technologies could help expand employment opportunities. Washington, DC: World Bank. Retrieved from http://www-wds.worldbank.org/ external/default/WDSContentServer/WDSP/IB/2013/09/09/000456286_20130909094 536/Rendered/PDF/809770WP0Conne00Box379814B00PUBLIC0.pdf. Accessed on October 23, 2015. Ranchordas, S. (2015). Does sharing mean caring? Regulating innovation in the sharing economy. Minnesota Journal of Law, Science & Technology, 16(1), 413–475. Ranchordás, S. (2016). Home sharing in the digital economy: The cases of Brussels, Stockholm, and Budapest. Brussels, Belgium: European Commission. Retrieved from https://ec.europa.eu/docsroom/documents/16950/attachments/1/translations/en/renditions/native. Accessed on May 2018. Rapkin, M. (2014, March). Uber cab confessions. GQ. Retrieved from http://www.gq.com/ story/uber-cab-confessions. Accessed on December 30, 2015. Rasmussen, S., & Kongshøj Madsen, P. (2016). “Denmark” in Nordic labour markets and the sharing economy. National Background Reports, Nordic Council of Ministers. Retrieved from http://dx.doi.org/10.6027/NA2017-904 Rating Legis. (2016). Impulse paper on the business authorisation/licencing requirements imposed both on peer providers and platforms: Barcelona, Berlin and Amsterdam. Brussels,

References    193 Belgium: European Commission. Retrieved from https://ec.europa.eu/docsroom/ documents/16948/attachments/1/translations/en/renditions/native. Accessed on May 2018. Rauch, D., & Schleicher, D. (2015). Like Uber, but for local government policy: The future of local regulation of the “shared economy.” Working Paper #21. New York, NY: Marron Institute of Urban Management. Ravenelle, A. (2016). Hustle: The lived experience of workers in the sharing economy. Ph.D. thesis, CUNY, New York, NY. Ravenelle, A. J. (2017). Sharing economy workers: Selling, not sharing. Cambridge Journal of Regions, Economy and Society, 10(2), 281–295. doi:10.1093/cjres/rsw043 Ravetz, J. (1990). The merger of knowledge with power: Essays in critical science. London: Mansell. Rawls, J. (1971). A Theory of Justice. Cambridge, MA: Harvard University Press. Resnick, P., & Zeckhauser, R. (2002). Trust among strangers in internet transactions: Empirical analysis of eBay’s reputation system. In Baye, M. R. (Ed.) The economics of the internet and e-commerce (Vol. 11, pp. 127–157). Bingley: Emerald Group Publishing Limited. RFS. (2015). 1099 Economy workforce report: Request for startup (RFS). Summary Report. Retrieved from http://www.slideshare.net/RequestsForStartups/the-2015-1099economy-workforce-report-preview. Accessed on November 15, 2015. Rheingold, H. (1994). The virtual community: Surfing the internet. London: Minerva. Richardson, L. (2015). Performing the sharing economy. Geoforum, 67, 121–129. doi:10.1016/j.geoforum.2015.11.004 Rizk, N. (2017, February 22). A glimpse into the sharing economy: An analysis of Uber driver-partners in Egypt. Retrieved from https://ssrn.com/abstract=2946083. Rochet, J.-C., & Tirole, J. (2003). Platform competition in two-sided markets. Journal of the European Economic Association, 1(4), 990–1029. doi:10.1162/154247603322493212 Rochet, J.-C., & Tirole, J. (2006). Two-sided markets: A progress report. Rand Journal of Economics, 37(3), 645–667. Roose, K. (2014a, September 18). Does Silicon Valley have a contract worker problem? New York Magazine. Retrieved from http://nymag.com/daily/intelligencer/2014/09/silicon-valleys-contract-worker-problem.html. Accessed on January 21, 2016. Roose, K. (2014b, April 24). The sharing economy isn’t about trust, it’s about desperation. New York Magazine. Retrieved from http://nymag.com/daily/intelligencer/2014/04/ sharing-economy-is-about-desperation.html. Accessed on May 30, 2015. Rosenblat, A., & Stark, L. (2015). Uber’s drivers: Information asymmetries and control in dynamic work. Retrieved from http://ssrn.com/abstract=2686227 Ross, J., Irani, L., Silberman, M., Zaldivar, A., & Tomlinson, B. (2010). Who are the crowdworkers? Shifting demographics in mechanical turk. Paper presented at the CHI ‘10 extended abstracts on human factors in computing systems, Atlanta, GA. Rossotto, C., Kuek, S., & Paradi-Guilford, C. (2012). New frontiers and opportunities in work. Washington, DC: World Bank. Retrieved from http://www-wds.worldbank.org/ external/default/WDSContentServer/WDSP/IB/2013/06/07/000356161_20130607151348/ Rendered/PDF/782660BRI0P1280Box0377336B00PUBLIC0.pdf. Accessed on October 23, 2015. Ruths, D., & Pfeffer, J. (2014). Social media for large studies of behavior. Science (New York, N.Y.), 346(6213), 1063–1064. doi: 10.1126/science.346.6213.1063 Rysman, M. (2009). The economics of two-sided markets. Journal of Economic Perspectives, 23(3), 125–143. doi:10.1257/jep.23.3.125 Sablik, T. (2014). The sharing economy. Are new online markets creating economic value or threatening consumer safety? Economics Focus, pp. 12–15. Saeedi, M., Shen, Z., & Sundaresan, N. (2015). The value of feedback: An analysis of reputation system. Retrieved from http://econ.ohio-state.edu/saeedi/noneg_sss.pdf Sanderson, I. (2011). Evidence-based policy or policy-based evidence? Reflections on Scottish experience. Evidence & Policy, 7(1), 59–76.

194   References Santi, P., Szell, M., Sobolevsky, S., Ratti, C., Resta, G., & Strogatz, S. H. (2014). Quantifying the benefits of vehicle pooling with shareability networks. Proceedings of the National Academy of Sciences of the United States of America, 111(37), 13290–13294. doi:10.1073/ pnas.1403657111 Sapir, A. (2005). Globalisation and the reform of European social models (Bruegel policy brief 25/01). Brussels, Belgium: Bruegel. Satzger, B., Psaier, H., Schall, D., & Dustdar, S. (2013). Auction-based crowdsourcing supporting skill management. Information Systems, 38(4), 547–560. doi:http://dx.doi. org/10.1016/j.is.2012.09.003 Schneiderman, E. (2014). Airbnb in the city. New York, NY: Office of the Attorney General of the State of New York. Retrieved from http://www.ag.ny.gov/pdfs/Airbnb report.pdf Scholz, T. (Ed.). (2013). Digital labor: The internet as playground and factory. London: Routledge. Scholz, T. (2016a). Uberworked and underpaid: How workers are disrupting the digital economy. Cambridge: Polity Press. Scholz, T. (2016b). Platform cooperativism: Challenging the corporate sharing economy. New York, NY: Rosa Luxemburg Stiftung. Retrieved from http://www.rosalux-nyc.org/ wpcontent/files_mf/scholz_platformcoop_5.9.2016.pdf Scholz, T., & Schneider, N. (2017). Ours to hack and to own: The rise of platform cooperativism, a new vision for the future of work and a fairer internet. New York, NY: OR Books. Retrieved from http://www.orbooks.com Schor, J. (2014). Debating the sharing economy. Boston, MA: Tellus Institute. Retrieved from http://greattransition.org/publication/debating-the-sharing-economy Schor, J. (2015). Getting sharing right. Contexts, 14(1), 14–15. Schor, J., & Attwood-Charles, W. (2017). The “sharing” economy: Labor, inequality, and social connection on for – profit platforms. Sociology Compass, 11(8), e12493. doi:10.1111/soc4.12493 Schor, J., & Fitzmaurice, C. (2015). Collaborating and connecting: The emergence of the sharing economy. In Reisch, L., & Thogersen, J. (Eds). Handbook of research on sustainable consumption (pp. 410–425). Cheltenham: Edward Elgar. Schor, J., Fitzmaurice, C., Carfagna, L., & Will-Attwood, C. (2014). Paradoxes of opennes and distinction in the sharing economy. Boston, MA: Boston College. Schrape, J.-F. (2017, October). Open-source projects as incubators of innovation: From Niche phenomenon to integral part of the industry. Convergence, 1–19. Schwartz, O. (2015, June 22). Humans pretending to be computers pretending to be humans. Retrieved from http://theliftedbrow.com/post/122116846059/excerpt-humans-pretending-to-be-computers. Accessed on December 7, 2015. Scot, M., & Eddy, M. (2014, September 2). Uber service banned across Germany by Frankfurt Court. New York Times. Retrieved from http://bits.blogs.nytimes.com/2014/09/02/uberbanned-across-germany-by-frankfurt-court/. Accessed on May 28, 2015. Selznick, P. (1948). Foundations of the theory of organization. American Sociological Review, 13(1), 25–35. Selznick, P. (1949). TVA and the grass roots; a study in the sociology of formal organization. Berkeley, CA: University of California Press. Sen, A. (2009). The Idea of Justice. London: Allen Lane. Seyedabrishami, S., Mamdoohi, A., Barzegar, A., & Hasanpour, S. (2012). Impact of carpooling on fuel saving in urban transportation: Case study of Tehran. Procedia – Social and Behavioral Sciences, 54(0), 323–331. doi:http://dx.doi.org/10.1016/ j.sbspro.2012.09.751 Shaw, A., Horton, J., & Chen, D. (2011). Designing incentives for inexpert human raters. Paper presented at the Proceedings of the ACM 2011 conference on computersupported cooperative work, Hangzhou, China. Shirky, C. (2010). Cognitive surplus: Creativity and generosity in a connected age. London: Penguin.

References    195 Shontell, A. (2011, December 11). My nightmare experience as a taskrabbit drone. Business Insider. Retrieved from http://www.businessinsider.com/confessions-of-a-taskrabbit-2011-12. Accessed on January 21, 2016. Silberman, M., & Irani, L. (2016). Operating an employer reputation system: Lessons from Turkopticon, 2008–2015. Comparative Labor Law & Policy Journal. Retrieved from http://ssrn.com/abstract=2729498. Accessed on March 20, 2016. Silberman, M., Irani, L., & Ross, J. (2010a). Ethics and tactics of professional crowdwork. XRDS, 17(2), 39–43. doi:10.1145/1869086.1869100 Silberman, M., Ross, J., Irani, L., & Tomlinson, B. (2010b). Sellers’ problems in human computation markets. Paper presented at the Proceedings of the ACM SIGKDD workshop on human computation, Washington, DC. Retrieved from http://dl.acm.org/ citation.cfm?doid=1837885.1837891 Silberzahn, R., Uhlmann, E. L., & Zhu, L. (2014). Pay as she goes: For stereotypically male jobs, women tend to be hired by the hour. Academy of Management Proceedings, 2014(1), 16273. Singer, N. (2014a, August 16). In the sharing economy, workers find both freedom and uncertainty. The New York Times. Retrieved from http://www.nytimes.com/2014/08/17/ technology/in-the-sharing-economy-workers-find-both-freedom-and-uncertainty.html. Accessed on November 30, 2015. Singer, N. (2014b, August 16). In the sharing economy, workers find both freedom and uncertainty. The New York Times. Retrieved from http://www.nytimes.com/2014/08/17/ technology/in-the-sharing-economy-workers-find-both-freedom-and-uncertainty.html. Accessed on May 30, 2015. Slee, T. (2015). What’s yours is mine: Against the sharing economy. New York, NY: OR Books. Smeaton, D. (2003). Self-employed workers: Calling the shots or hesitant independents? A consideration of the trends. Work, Employment and Society, 17(2), 379–391. doi:10.1177/0950017003017002008 Smith, R., & Leberstein, S. (2015). Rights on demand: Ensuring workplace standards and worker security in the on-demand economy. Washington, DC: National Employment Law Project (NELP). Smorto, G. (2016). Impulse paper No. 02 on the business authorisation/licencing requirements imposed both on peer-providers and platforms in the accommodation/tourism sector in Paris, Rome, Milan and London. Brussels, Belgium: European Commission. Retrieved from https://ec.europa.eu/docsroom/documents/16949/attachments/1/translations/en/ renditions/native. Accessed on May 2018. Smyrnaios, N. (2017). Les GAFAM contre l’internet: Une économie politique du numérique. France: InaGlobal. Smythe, D. W. (1977). Communications: Blindspot of Western Marxism. CTheory, 1(3), 1–27 Snow, D., & Benford, R. (1988). Ideology, frame resonance, and participant mobilization. International Social Movement Research, 1, 197–217. Snow, D., Rochford, E., Worden, S., & Benford, R. (1986). Frame alignment processes, micromobilization, and movement participation. American Sociological Review, 51(4), 464–481. doi:10.2307/2095581 Solow, R. M. (1990). The labor market as a social institution. Oxford: Basil Blackwell. Spasova, S., Bouget, D., Ghailani, D., & Vanhercke, B. (2017). Access to social protection for people working on non-standard contracts and as self-employed in Europe: A study of national policies. Brussels, Belgium: European Commission, Directorate General for Employment, Social Affairs and Inclusion. Sperling, G. (2015). How Airbnb combats middle class income stagnation: Airbnb. Retrieved from https://airbnb.box.com/shared/static/jrry0if4cgjrnvw1ykrpmw8ugivrx2mx.pdf Accessed on January 31, 2016. Sprague, R. (2015). Worker (mis)classification in the sharing economy: Square pegs trying to fit in round holes. Journal of Labor & Employment Law. Retrieved from http://ssrn.com/ abstract=2606600. Accessed on December 2, 2015.

196   References Srnicek, N. (2017). Platform capitalism. Cambridge: Polity Press. Standing, G. (2011). The Precariat: The dangerous class. London: Bloomsbury. Stanton, C., & Thomas, C. (2014). Landing the first job: The value of intermediaries in online hiring. London: Centre for Economic Performance, London School of Economics and Political Science. Stewart, A., & Stanford, J. (2017). Regulating work in the gig economy: What are the options? The Economic and Labour Relations Review, 28(3), 420–437. doi:10.1177/1035304617722461 Stigler, G. (1971). The theory of economic regulation. Bell Journal of Economics, 2(1), 3–21. Stokes, K., Clarence, E., Anderson, L., & Rinne, A. (2014). Making sense of UK collaborative economy. London: NESTA. Retrieved from http://www.nesta.org.uk/sites/default/ files/making_sense_of_the_uk_collaborative_economy_14.pdf Stolle, D. (2002). Trusting strangers: The concept of generalized trust in perspective. OZP State Institute of Science and Politics, 31(4), 397–412. Stolle, D., & Hooghe, M. (2004). The roots of social capital: Attitudinal and network mechanisms in the relation between youth and adult indicators of social capital. Acta Politica, 39(4), 422. doi:10.1057/palgrave.ap.5500081 Strassheim, H., & Kettunen, P. (2014). When does evidence-based policy turn into policybased evidence configurations, contexts and mechanisms. Evidence and Policy, 10(2), 259–277. doi:10.1332/174426514X13990433991320 Strom, S., & Schmitt, M. (2016). Protecting Workers in a Patchwork Economy. Washington D.C.: The Century Foundation. Retrieved from: https://tcf.org/content/report/protecting-workers-patchwork-economy/. Accessed on April 16, 2016. Summers, L., & Balls, E. (2015). Report of the commission on inclusive prosperity. Washington, DC: Center for American Progress. Retrieved from https://cdn.americanprogress.org/wp-content/uploads/2015/01/IPC-PDF-full.pdf Sundararajan, A. (2014). Peer-to-peer businesses and the sharing (collaborative) economy: Overview, economic effects and regulatory issues. Written Testimony for the Hearing Title, the Power Of Connection: Peer-to-Peer Businesses, held by the Committee on Small Businesses of the US House of Representatives, September 15, 2014. Sundararajan, A. (2016). The sharing economy: The end of employment and the rise of crowd-based capitalism. Cambridge, MA: MIT Press. Sunil, J., & Noah, Z. (2015). Policymaking for the sharing economy: Beyond whack-a-mole. Toronto, Canada: Mowat Centre, University of Toronto. Retrieved from http://mowatcentre.ca/wp-content/uploads/publications/106_PolicymakingForTheSharingEconomy.pdf Surowiecki, J. (2004). The wisdom of crowds: Why the many are smarter than the few and how collective wisdom shapes business, economies and societies. New York, NY: Doubleday. Swamynathan, G., Wilson, C., Boe, B., Almeroth, K., & Zhao, B. Y. (2008). Do social networks improve e-commerce? A study on social marketplaces. In C. Faloutsos, T. Karagiannis, & P. Rodriguez (Eds.), WOSN ‘08 (pp. 1–6). New York, NY: ACM. Tadelis, S. (2016). Reputation and feedback systems in online platform markets. Annual Review of Economics, 8(1), 321–340. doi:10.1146/annurev-economics-080315-015325 Taylor, P., & Jordan, T. (2004). Hacktivism and cyberwars: Rebels with a cause? London: Routlege. Taylor, T. (2015, May 28). The sharing economy. Retrieved from http://conversableeconomist.blogspot.com.es/2015/05/the-sharing-economy.html. Accessed on October 20, 2015. Teodoro, R., Ozturk, P., Naaman, M., Mason, W., & Lindqvist, J. (2014). The motivations and experiences of the on-demand mobile workforce. Paper presented at the Proceedings of the 17th ACM conference on computer-supported cooperative work & social computing, Baltimore, MD. Terranova, T. (2013 [2000]). Free labor. In T. Scholz (Ed.), Digital labor: the internet as playground and factory (pp. 33–57). New York, NY: Routledge. Thebault-Spieker, J., Terveen, L., & Hecht, B. (2015). Avoiding the south side and the suburbs: The geography of mobile crowdsourcing markets. Paper presented at the

References    197 Proceedings of the 18th ACM conference on computer-supported cooperative work & social computing, Vancouver, BC, Canada. Thelen, K. (2014). Varieties of liberalization and the new politics of social solidarity. New York, NY: Cambridge University Press. Thierer, A., Koopman, C., Hobson, A., & Kuiper, C. (2015). How the internet, the sharing economy, and reputational feedback mechanisms solve the “lemons problem.” Retrieved from http://ssrn.com/abstract=2610255 Thorngate, W. (2001). The social psychology of policy analysis. Journal of Comparative Policy Analysis: Research and Practice, 3(1), 85–112. doi:10.1080/13876980108412655 Todolí-Signes, A. (2017). The “gig economy”: Employee, self-employed or the need for a special employment regulation? Transfer: European Review of Labour and Research, 23(2), 193–205. doi:10.1177/1024258917701381 Torriti, J. (2010). Impact assessment and the liberalization of the EU energy markets: Evidence-based policy-making or policy-based evidence-making? Journal of Common Market Studies, 48(4), 1065–1081. doi:10.1111/j.1468-5965.2010.02089.x Trottier, D., & Fuchs, C. (Eds.). (2014). Social media, politics and the state: Protests, revolutions, riots, crime and policing in the age of Facebook, Twitter and YouTube. London: Routledge. Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of choice. Science, 211(4481), 453. doi:10.1126/science.7455683 Uber. (2014). Uber’s economic impact on the city of Chicago. Retrieved from http://newsroom.uber.com/chicago/2014/03/uberdata-ubers-economic-impact-on-the-city-of-chicago/. Accessed on June 8, 2015. Uber. (2015a). DUI rates decline in Uber cities. Retrieved from http://newsroom.uber.com/ chicago/2014/05/dui-rates-decline-in-uber-cities/. Accessed on June 8, 2015. Uber. (2015b). Improving reliability on military bases. Retrieved from http://newsroom.uber. com/sd/2015/04/uberimpact-better-serving-those-who-serve-us/. Accessed on June 8, 2015. Uber. (2015c). Our 2015 #Uber impact partners! Retrieved from http://newsroom.uber.com/ seattle/2014/12/our-2015-uberimpact-partners/. Accessed on June 8, 2015. Uber. (2015d). Uber’s impact in Edmonton. Retrieved from http://newsroom.uber.com/ edmonton/2015/03/ubers-impact-in-edmonton/. Accessed on June 8, 2015. Uber. (2015e). Uber’s impact on Kansas City. Retrieved from http://newsroom.uber.com/ kc/2015/03/ubers-impact-in-kansas-city/. Accessed on June 8, 2015. Uber. (2015f). Ubernomics: How ridesharing can impact the German economy (part I). Retrieved from http://newsroom.uber.com/berlin/en/2015/03/diwdice-studie-digitalisierung-bietet-grose-chancen-fur-stadtische-mobilitat-veraltete-regulierung-verhindert-vorteile-fur-kunden-fahrer-und-kommunen-2/. Accessed on June 8, 2015. Uber. (2015g). Ubernomics: How ridesharing can impact the German economy (part II). Retrieved from http://newsroom.uber.com/berlin/en/2015/03/diwdice-studie-digitalisierung-bietet-grose-chancen-fur-stadtische-mobilitat-veraltete-regulierung-verhindert-vorteile-fur-kunden-fahrer-und-kommunen-2-2-2/. Accessed on June 8, 2015. Uber. (2018). White paper on work and social protection in Europe. Retrieved from https:// newsroomadmin.uberinternal.com/wp-content/uploads/2018/02/Uber-White-Paperon-Work-and-Social-Protections-in-Europe.pdf. Accessed on May 2018. Uddin, Z. (2012, September 11). The dystopian digital sweatshop that makes the internet run. Retrieved from http://www.alternet.org/labor/dystopian-digital-sweatshop-makesinternet-run. Accessed on December 1, 2015. Uhlmann, E. L., & Silberzahn, R. (2014). Conformity under uncertainty: Reliance on gender stereotypes in online hiring decisions. Behavioral and Brain Sciences, 37(1), 103–104. doi:10.1017/S0140525X13001921 UNDESA. (2014). World urbanization prospects: The 2014 revision, highlights. New York, NY: United Nations, Department of Economic and Social Affairs (UNDESA). Retrieved from http://esa.un.org/unpd/wup/Highlights/WUP2014-Highlights.pdf. Accessed on August 19, 2015.

198   References Urtasun, A., & Núñez, I. (2012). Work-based competences and careers prospects: A study of Spanish employees. Personnel Review, 41(4), 428–449. doi:10.1108/00483481211229366 Vakharia, D., & Lease, M. (2013). Beyond AMT: An analysis of crowd work platforms. Retrieved from https://arxiv.org/pdf/1310.1672.pdf Valenduc, G., & Vendramin, P. (2016). Work in the digital economy: Sorting the old from the new. Brussels, Belgium: ETUI (European Trade Union Institute). Retrieved from https://www.etui.org/Publications2/Working-Papers/Work-in-the-digital-economysorting-the-old-from-the-new Van de Donk, W., Loader, Br., Nixon, P., & Rucht, D. (Eds.). (2004). Cyberprotest: New media, citizens and social movements. London: Routledge. Van de Glind, P. (2013). The consumer potential of collaborative consumption. Identifying (the) motives of Dutch collaborative consumers & measuring the consumer potential of collaborative consumption within the municipality of Amsterdam. M.Sc. thesis, Utrecht University, Utrect. Retrieved from http://dspace.library.uu.nl/handle/1874/280661. Accessed on July 12, 2015. Van Doorn, N. (2017). Platform labor: On the gendered and racialized exploitation of low-income service work in the “on-demand” economy. Information, Communication & Society, 20(6), 898–914 (AoIR special issue). Van Lancker, W. (2012). The European world of temporary employment: Gendered and poor? European Societies, 14(1), 83–111. doi:10.1080/14616696.2011.638082 Vasagar, J. (2014, April 17). Uber taxi service suffers setback in Berlin. Financial Times. Retrieved from http://www.ft.com/intl/cms/s/0/1591faf2-c638-11e3-ba0e-00144feabdc0. html-axzz3bRKe8rMn. Accessed on May 28, 2015. Vaughan, R., & Daverio, R. (2016). Assessing the size and presence of the collaborative economy in Europe. London: PWC. Vaughan, R., & Hawksworth, J. (2014). The sharing economy: How will it disrupt your business? Megatrends: The Collisions. PwC Report. Vedung, E. (2010). Four waves of evaluation diffusion. Evaluation, 16(3), 263–277. Virtanen, M., Kivimäki, M., Joensuu, M., Virtanen, P., Elovainio, M., & Vahtera, J. (2005). Temporary employment and health: a review. International Journal of Epidemiology, 34(3), 610–622. VVA & Spark. (2018). Study on the assessment of the regulatory aspects affecting the collaborative economy in the tourism accommodation sector in the 28 member states. Task 2 Report. Brussels, Belgium: European Commission. Retrieved from https://publications.europa.eu/en/publication-detail/-/publication/c7a7b5bb-526d-11e8-be1d01aa75ed71a1/language-en/format-PDF/source-70300102. Accessed on May 2018. Walker, E. (2015). Beyond the rhetoric of the sharing economy. Contexts, 14(1), 15–17. Wallsten, S. (2015). The competitive effects of the sharing economy: How is Uber changing taxis? New York, NY: Technology Policy Institute. Weber, L., & Silverman, R. (2015a, January 28). On­demand workers: “We are not robots” – Is technology liberating or squeezing the new class of freelance labor? The Wall Street Journal. Retrieved from http://www.wsj.com/articles/on-demand-workers-we-are-notrobots-1422406524. Accessed on November 30, 2015. Weber, L., & Silverman, R. (2015b, January 28). On­demand workers: “We are not robots” – Is technology liberating or squeezing the new class of freelance labor? The Wall Street Journal. Retrieved from http://www.wsj.com/articles/on-demand-workers-we-are-notrobots-1422406524. Accessed on May 31, 2015. Weiner, J. (2015a, February 20). The hidden costs of being an Uber driver. The Washington Post. Retrieved from http://www.washingtonpost.com/business/get-there/the-hiddencosts-of-being-an-uber-driver/2015/02/20/22a0c49f-d829-4368-b67e-fb9d049badb5_ story.html. Accessed on November 28, 2015. Weyl, E. G. (2010). A price theory of multi-sided platforms. American Economic Review, 100(4), 1642–1672.

References    199 Weiner, J. (2015b, February 20). The hidden costs of being an Uber driver. The Washington Post. Retrieved from http://www.washingtonpost.com/business/get-there/the-hiddencosts-of-being-an-uber-driver/2015/02/20/22a0c49f-d829-4368-b67e-fb9d049badb5_ story.html. Accessed on May 28, 2015. Wittel, A. (2011). Qualities of sharing and their transformation in the digital age. International Review of Information Ethics, 15, 3–8. Wood, A. J., Graham, M., & Lehdonvirta, V. (2016). Virtual production networks: Fixing commodification and disembeddedness. Proceedings of the internet, policy & politics conference, Oxford Internet Institute, University of Oxford, Oxford, UK, 22 September. World Economic Forum (WEF). (2013). Young global leaders sharing economy dialogue position. Paper of World Economic Forum (WEF). Retrieved from http://www3.weforum.org/docs/WEF_YGL_CircularEconomyInnovation_PositionPaper_2013.pdf World Economic Forum (WEF). (2014). Towards the circular economy: Accelerating the scale-up across global supply chains. Retrieved from http://www3.weforum.org/docs/ WEF_ENV_TowardsCircularEconomy_Report_2014.pdf Wosskow, D. (2014). Unlocking the sharing economy. An independent review. London: UK Department for Business, Innovation and Skills. Retrieved from https://www.gov.uk/ government/uploads/system/uploads/attachment_data/file/378291/bis-14-1227-unlocking-the-sharing-economy-an-independent-review.pdf Wright, J. (2004). One-sided logic in two-sided markets. Review of Network Economics, 3(1), 44–64. Yamagishi, T., Cook, K., & Watabe, M. (1998). Uncertainty, trust, and commitment formation in the United States and Japan. American Journal of Sociology, 104(1), 165–165. Yamagishi, T., & Yamagishi, M. (1994). Trust and commitment in the United States and Japan. Motivation and Emotion, 18(2), 129–166. doi:10.1007/BF02249397 Zervas, G., Proserpio, D., & Byers, J. (2014). The rise of the sharing economy: Estimating the impact of Airbnb on the hotel industry. Research Paper. Boston, MA: Boston U. School of Management. Zervas, G., Proserpio, D., & Byers, J. (2015). A first look at online reputation on Airbnb, where every stay is above average. Retrieved from http://ssrn.com/abstract=2554500 Zimmermann, J. (2015, March 12). This is what it takes to make $2,000 a week working on taskrabbit. Money. Retrieved from http://time.com/money/3714829/working-fortaskrabbit/. Accessed on November 30, 2015. Zittrain, J. (2008). Ubiquitous human computing. Philosophical Transactions. Series A: Mathematical, Physical, and Engineering Sciences, 366(1881), 3813. doi:10.1098/ rsta.2008.0116 Zittrain, J. (2009, December 7). The internet creates a new kind of sweatshop. Newsweek. Retrieved from http://www.newsweek.com/internet-creates-new-kind-sweatshop-75751. Accessed on December 1, 2015. Zrenner, A. (2015). The ethics of regulating the sharing economy. Durham, NC: The Kenan Institute for Ethics, Duke University. Zumbrun, J., & Sussman, A. (2015, July 26). Proof of a “gig economy revolution” is hard to find. The Wall Street Journal. Retrieved from http://www.wsj.com/articles/proof-ofa-gig-economy-revolution-is-hard-to-find-1437932539. Accessed on January 21, 2016.

This page intentionally left blank

Index Note: Page nos. in bold refers to table and italics refers to figures Advocate General (AG), 115–116 affiliation, 21 aggregate welfare effects, 106–107 Airbnb, 25, 25n4, 26, 29, 31, 32, 36, 39, 44, 45, 51–53, 57–60, 63–70, 68n24, 135–139, 146, 153–154, 159, 165 algocracy, 96 algorithm, 96–97 algorithmic labour, 125 algorithm management, 168 Amazon Mechanical Turk, 44 Application Programme Interface (API), 79 attention economy, 129 audience-makers, 23 automated workers control, 95–97 autonomy, 154 Barcelona-based crowdfunding platform, 139–140 bartering, lending, renting, gifting and swapping, 39 Behavioral and Brain Sciences, 104 big data opportunity, 167 BlaBlaCar, 36, 58 boundaries, 151–152 Bring-to-Market (BTM) cost, 37–38, 63 business optimism and economic laissez faire, 50 business-to-consumer (B2C), 41 car sharing vs. ride services, 26–27, 27 centralisation vs. decentralisation, 25–26 Chartered Institute for Personnel and Development (CIPD), 85, 89–91

cognitive biases, 71 collaborative consumption, 39 collaborative economy, 40, 46 Commission Eurobarometer, 121 The Common Good: Ethics and Rights in Cybersecurity, 132 competition law, 71 competition law potential implications, 160 Confederation of Danish Trade Unions, 35 congestion, 24 consumer protection, 37 consumer welfare and distributional effects, 62–65 control and cost trade-off, 33, 33–34 conversation, 136 co-operativism, 144–148 corporate social responsibility (CSR), 127 CouchSurfing, 29, 36, 51, 56, 67 counterparts, 28 Court of Justice of the European Union (CJEU), 115 cross-border interest, 127 cross-group externalities definition, 21 Crowdflower, 88–93, 156 crowdfunding platform, 125 crowdsourcing platform, 125 cultural capital, 55n23 data and privacy, 71, 160 data collection, 132 debate, 117–122 Delphi study, 37–38n3 demand coordinators, 23 digital activism scholarship, 131

202   Index digital economy, 124 digital labour markets, 41, 73–76, 107–110, 128, 152–153 conceptualisation and dimensions, 76–85 efficiency, 100–107 employment status, 91 motivations, 90–91 NSW and social protection, 97, 98, 99, 99–100 regulation and policy, 111–122 rhetorics and economic hypotheses, 85–87 socio-demographic profile, 87–90 working conditions, 91–97 digital labour organisation, 131 digital platforms, 17, 19–20, 24, 25, 27, 31, 41 digital political economy, 129 digital sharing, 48 digital transformation, 162–163 discourse, 154 discrimination, 164 distributional and stratification effects, 57–58 distributional effects, 168 earnings, 91–93 eBay, 29 economic classification, 41 economic crisis, 37–38 economic literature, 19, 76 economic theory, 161 economies literature, 131 efficiency, market, 100–107 Elance-oDesk, 88, 103 empirical evidence, 39, 153, 164 empirical research, 24–25 Employment and Social Affairs, 35 employment status, 91 encadrement, 150 environmental impacts, 60–61 equally inconclusive, 154 ESRC project, 132 EU agenda on collaborative economy, 120

EU-level context and developments, 119 EU-level flexibilisation of work, 119–120 European Agency for Safety and Health at Work, 35 European Central Bank, 140 European Commission, 40, 46, 68n24, 68n25, 97, 120–122 European Economic and Social Committee (EESC), 162 European framework, 127 European Parliament, 46, 120–122 Europe, legal disputes, 115–116 evidence, 153–156, 164–165 ex ante, 30, 59, 163 experimental regulation, 117 experimenter as employer framework, 102 ex post, 163 fair and dignified support infrastructure (FDSI), 163–164 Fair Labor Standard Act (FLSA), 111 fear of retaliation, 29 Federal Trade Commission, 26 Federal Trade Commission Relay Rides, 42n9 Financial Times, 58 first-order effects, 60, 154 flexibility, 113, 154–155 floating signifiers, 40, 152 for-profit vs. non-profit, 137 Foundations, Discourses and Limits of the Collaborative Economy: An Exploratory Research, 132 Freecycle, 36 Freelancers, 80, 80n4 French Environment and Energy Management Agency (ADEME), 61 Greek social movements, 140 greenhouse gas (GHG) emission, 60–61 growth, 113 Harvard Business Review, 52, 114 heterogeneity, 24–25

Index    203 heterosexual dating club, 23 heuristic conceptual mapping, 43, 43–44 Homo Oeconomicus, 104 Human Intelligence Tasks (HIT), 79 hypercapitalism, 124 identification (ID) checks, 160 identification system, 28 ideological production, 123–126 challenges, 134–139 commons, 139–144 discourses, 139–144 integrated theoretical framework, 126–131 methodology, 131–134 spectrum outlook, 144–148 iLabour index, 127 illusions of engagement, 129 information asymmetries, 71 and cognitive biases, 159–160 information asymmetry, 94 Information Communication Technology (ICT), 108 inherently frictional markets, 24–25 Instacart, 113 integrated theoretical framework, 126–131 intensional definition, 17n1 intentional collusive behaviour, 29 Internal Market Committee, 120 International Monetary Fund, 140 Internet, 30, 157 job opportunities, 163 labour-intensive services, 83 labour issues, 163–164 labour platforms, 35–37 consumer welfare and distributional effects, 62–65 distributional and stratification effects, 57–58 environmental impacts, 60–61 lobbying, 51–54 ratings, 65–67

regulation and policy, 68–71 rhetorical and discourse analysis, 48–51 sharing economy, 45–48 social capital and motivation, 54–57 socio-economic impacts, 61–62 trajectory and conceptual issues, 37–45 labour-specific issues, 161–162 lack of evidence, 161 legal disputes in Europe, 115–116 liability and insurance, 71, 159 liability insurance, 164 liability rules, 37 licencing and certification schemes, 71, 160 litigation, 111–114, 112 lobbying, 51–54, 164–168, 165 long tail effects, 104–105 Lyft, 26–27, 95, 156 litigation, 111–112 manual labour work, 78 market efficiency, 100–107 market frictions, 101–103 market-makers, 23 media accounts, 85 medium-term research project, 166 micro-entrepreneurs, 58 micro-entrepreneurship, 85 micro-tasking, 78–79, 87–88 minimum wage, 164 Mobile Crowd, 89 mobile labour markets (MLMs), 75–77, 77, 154 distributional effects, 107 interactive services, 83–85 physical services, 82–83, 89–90 moral hazard, 94 moral valence, 94 motivations, 90–91 MTurk, 78–79, 87–95, 108, 156 multi-sided platforms (MSPs), 18–22, 22, 42 Munchery, 113 MyClean, 113

204   Index National Endowment for Science, Technology and the Arts (NESTA), 47 negative externalities, 71, 159 neo-liberal co-optation, 51 net aggregate effects, 106–107 NFP-sharing and cooperation, 39 no need for indirect network effect, 21 non-profit organisations, 137 non-standard work (NSW), 73–74, 97, 98, 99, 99–100, 109–110, 156 not-for-profit (NFP) platforms, 40 Nuit Debout mobilisation, 141 occupational safety and health risk, 78 oDesk, 25, 25n4, 80–81n4, 88, 100–103 Office for National Statistics (ONS), 45 on-demand labour trade, 39 on-demand versus scheduled transactions, 26–27 one-way indirect network effect, 20 online crowdlabour markets, 95 online exchange, 28 online labour markets (OLMs), 75–77, 77, 86, 100–102, 105–107, 155, 167–168 micro-tasking, 78–79, 87–88 tasking, 80–81, 88–89 Open Food Network, 142–143, 147 operating systems, 23 opportunity and access, 113 ostensive definition, 17n1 participants, 131–134, 132–133 passenger assignment, 96 pass-through, 20n3 payment systems, 23 peers-to-businesses (P2Bs), 41, 44 peer-to-peer (P2P) dimension, 18, 41, 42, 44, 45, 48, 60, 62–65 peer-to-peer Internet platforms, 40 person-to-person carsharing platform, 26 pin money and flexers, 154 platform capitalism, 124–125

platform cooperativism, 128–131 platformisation scholarship, 131 platform labour, 125 policy and regulation, 68–71, 157–164 policy makers, 162 political activation, 152 price, 25 pricing mechanism, 25, 66 production cycle, 125 product service systems, 39 protection of privacy, 164 public interest, 52 public policy, 134 public relations (PR), 127 PwC, 46, 46n13 quality of object, 28 ratings, 65–67 and governance, 27–31 redistribution markets, 39 regulation and policy, 68–71 Relay Rides, 26, 26n7, 27, 44 reputational ratings, 66 reputation rating, 28 rhetorical and discourse analysis, 48–51 rhetorical discourses, 86 rhetorical themes, 152–153 rhetoric and movement, 51 robust evidence, 161 Routine Biased Technical Change (RBTC), 78 science, role of, 165, 165 security, 113 self-contained project, 78 self-defined sharing platforms, 36 self-define practice, 40 self-defining practices, 152 self-regulation, 65–67 sharing economy, 31–34, 39–41, 45–48, 125, 127 classifications and determinants, 22–27 definition, 17–18 ratings and governance, 27–31 two-sided/multi-sided markets, 18–22

Index    205 sharing movement, 152 short-research project, 166 single market relevance, 127 size determinants, 23–24, 24 Skill Biased Technological Change (SBTC) hypothesis, 108 social capital and motivation, 54–55n22, 54–57 social interaction, 29 social norms and values, 163 social pessimism, 50 social process, 124 social protection, 97, 98, 99, 99–100 and health insurance, 164 social utopianism, 49–50n15, 50, 152 society, impact on, 128 socio-cultural trends, 38 socio-demographic dimensions, 97 socio-demographic profile, 87–90 socio-demographic profiles, 168 socio-economic drivers, 39 socio-economic impacts, 61–62 socio-economic status (SES), 107 socio-economic structure, 125 software platforms, 23 Spanish law, 134 spectrum outlook, 144–148 State and capital, 125 superstar effects, 104–105, 104–105n15 sustainability, 55 TaskRabbit, 24–25, 25n4, 26, 27, 36, 44, 66, 82–83, 82n5, 89, 92, 103, 107, 112, 154 taxation, 70, 159 techno-optimism, 140 1099 Economy Workforce Report, 93 Time Banks, 36, 36n2 trade-off, 33, 33–34 transaction cost, 22–23, 66

transferable sharing right (TSR) mechanism, 158 TripAdvisor, 29, 67 true and pseudo-sharing, 39 trust and reputation systems, 27–28, 27n8 trust transitivity, 28 Turk, 107–108 Turker Nation, 94 Turkopticon, 94 TV market, 20 two-side markets (2SMs), 18–22, 22, 41, 42 two-way indirect network effect, 20 typology mapping, 17–18 Uber, 25–27, 27, 32, 33, 36, 39, 44–45, 47, 52–54, 58–59, 65–70, 95, 96–97, 111, 153–154, 156, 165 litigation, 111–114 San Francisco (SF), 96 in Spain, 134–139 Uber Pool, 135 uncertainties, 28 uncertainty, 164–165 United States legal disputes, 111–114, 112 Upwork, 80–81, 80n4, 96 U-shaped job polarisation, 78 verification system, 28 virtual production network, 127 wired labour, 108 working conditions, 91–97 work–life balance, 154 World Economic Forum (WEF), 38n4 Zipcar, 31, 41, 51, 56 Zombie capitalism, 124