“I JUST WANT TO SURVIVE.” A comparative study of food courier riders in three African cities Edward Webster and Fikile Masikane December 2021 Trade Union Competence Centre for Sub-Saharan Africa

ABSTRACT A new worker has emerged in the digital economy. This worker is subject to a new business model based on a form of algorithmic management. Promising freedom, flexibility, self-employment and Shared Business Ownership(SBO), these technology companies are opening new economic opportunities in Africa. However, in our comparative study of food courier motorbike riders in Accra(Ghana), Johannesburg(South Africa) and Nairobi(Kenya), a different picture emerges. A new form of precarious work is evident. It consists of predominantly men, working long hours with no paid leave or social security benefits. Riders face constant risks of accidents and harassment, and the companies they work for do not provide any basic occupational health and safety protection. Algorithmic management is turning out to be a doubleedged sword. On the one hand, it is leading to an extension of authoritarian managerial control over workers, by increasing their insecurity and by dramatically deepening inequality globally between extremely wealthy senior managers and a growing reserve of precarious workers. On the other hand, by technologically linking riders together, companies have increased workers’ workplace bargaining power. This provides riders with the opportunity to develop collective solidarity and even conduct strike action. Through our research, we identified union-like hybrid collectives offering mutual aid. This opens the possibility of riders being able to exercise structural, associational and societal power. THE GOAL OF OUR RESEARCH WAS TWO-FOLD: 1. To understand the working lives of food courier motorbike riders 2. To use our findings to advocate for policy change by facilitating the self-organisation of riders. Throughout our study we used participatory action research. Photograph 1: Food courier motorbike riders at Campus Square, Auckland Park, Johannesburg, South Africa(March 2020) by Fikile Masikane. 3

PHOTOGRAPHS All black and white photos used are part of the James Oatway Photo Exhibition. Photograph 1: Food courier motorbike riders at Campus Square, Auckland Park, Johannesburg, South Africa(March 2020) by Fikile Masikane. Photograph 2:“We have to wear protective gear. There are many accidents in this line of work.” Riders’ Photo Competition, Hillbrow, Johannesburg, South Africa (September 2021) by Justice Tafamba. Photograph 3: ID photograph submitted with other required documents to Uber Eats(May 2021) by Fikile Masikane. Photograph 4:“I have to carry this[pepper spray] when I do deliveries at night.” Riders’ Photo Competition, Yeoville, Johannesburg, South Africa(September 2021) by Cebile Mkhabela. Photograph 5:“Attending a wedding of a fellow friend and food courier rider.” Riders’ Photo Competition, Orange Grove, Johannesburg, South Africa(September 2021) by Justice Tafamba. Photograph 6:“Praying for God’s protection from accidents.” Riders’ Photo Competition, Hillbrow, Johannesburg, South Africa(September 2021) by Justice Tafamba. DIAGRAMS Photograph 7: The rules and regulations of the Brothers of Melville(BOM), Campus Square, Johannesburg, South Africa (March 2020) by Fikile Masikane. Photograph 8:“Our leader addressing us in a meeting.” Riders’ Photo Competition, Parktown, Johannesburg, South Africa(September 2021) by Justice Tafamba. Photograph 9: Food courier riders’ bikes parked at the Pieter Roos Park, Parktown, Johannesburg, South Africa (April 2021) by Fikile Masikane. Photograph 10: Londi Kumalo, one of the few women food courier riders interacting with the players at the soccer match at Pieter Roos Park, Parktown, Johannesburg, South Africa (April 2021) by Fikile Masikane. Photograph 11: The riders being divided into groups for the final game. Pieter Roos Park, Parktown, Johannesburg, South Africa(April 2021) by Fikile Masikane. Photograph 12: McDonald riders vs Uber Eats riders at a soccer tournament. Riders’ Photo Competition, Pieter Roos Park, Parktown, Johannesburg, South Africa(September 2021) by Justice Tafamba. TABLES Diagram 1: Tassinari and Maccarrone’s(2020) four-sided model of food delivery platforms Diagram 2: Africa’s digital platforms Diagram 3: Subsea cables across the world providing access to digital devices Diagram 4: Hybrid forms of worker organisations Diagram 5: Trade union power resources GRAPHS Graph 1: Access to mobile technologies Graph 2: Riders’ ages by city and country Graph 3: Riders’ marital status by city and country Graph 4: Riders’ accommodation: rent vs. own Graph 5: Riders’ length of employment as courier riders Graph 6: Riders’ perception of employment relationship: a worker or self-employed partner? Graph 7: What attracted riders to work as food courier riders? Graph 8: Is being a food courier rider, workers’ only occupation? Table 1: Who do you ride for? Table 2: Estimated costs of buying a bike Table 3: Estimated costs of servicing a bike Table 4: In a relationship between clients and yourself as a rider, how do you perceive the company’s position? Graph 9: Riders who owned vs. rented their motorbikes Graph 10: How frequently do riders get their bikes serviced? Graph 11: The nature of the relationship between the riders, the courier company, and the client Graph 12: Riders’ knowledge about how much commission they were being charged Graph 13: Impact of riders’ working hours on family time Graph 14: Workplace problems riders face Graph 15: Where is your daily place of work? Graph 16: Why riders join WhatsApp groups Graph 17: Rider strikes: What were the demands? 4

ABBREVIATIONS AND ACRONYMS BOM COTU-K CCMA ETUI FES GHS/ GH₵ IFAT ILO ITF KES/ KSh KPIs LRA Brothers of Melville Central Organization of Trade Unions Kenya Commission for Conciliation, Mediation and Arbitration European Trade Union Institute Friedrich-Ebert-Stiftung Ghanaian Cedi Indian Federation of App-Based Transport Workers International Labour Organization International Transport Federation Kenyan Shillings Key Performance Indicators Labour Relations Act NUPSAW PBs PRA PPE SACCO SADA SBO SCIS SEWA TAWU-K TUCC UU 5 National Union of Public Service and Allied Workers Personal Bests Power Resources Approach Personal Protective Equipment Savings and Credit Cooperative Organization South Africa in the Digital Age Shared Business Ownership Southern Centre for Inequality Studies Self-Employed Women’s Association Transport and Allied Workers Union of Kenya Trade Union Competence Centre United Ugandans

TABLE OF CONTENTS 1. ABSTRACT 1. EXECUTIVE SUMMARY 2. INTRODUCTION Prelude: A day in the working life of two food courier riders 2. INTRODUCTION Background The rise of platform capital Authoritarian Management? Purpose of the study A double-edged sword Research methodology Africa’s platform economy Food couriers enter Africa 3. UBER ECONOMICS Fo A o N d D co L u A ri B er O c U o R m R pa E n L y AT re IO ve N n S u H e IPS Ma 3 rk . e R t E e S n E t A ry R s C tr H at M eg E y THODS Driver attractTihoen cahnadllerentgeentoiofncosntrdautecgtiyng ethnographic research during the COVID-19 pandemic The law of dRimesineiasrhcihngsteramtepgloyyment relations TheSepeeirnilgs othfeidrliedecrasp’ awcoitryld of work through their own lenses The perils ofAincdtieopnernedseanrcceh: the seeds of a rider organisation Con 4 c . l R u E si S o E n ARCH FINDINGS PART 1: WHO ARE THE FOOD COURIER RIDERS? A DEMOGRAPHIC PROFILE 4. UBER CAREERS: N R E e T g W ist O e R rin K g A a C s C a E f S o S o , d E c X o P u E r C ie T r A ri T d I e O r N in S S A o N ut D h C A O fr N ica S , T b R y A F IN ik T ile S Masikane Nationality of riders Demographic profile and network access Employment history Food courier riders overwhelmingly men Education attained Financial and non-financial constraints Riders’ marital status Conclusion Rent or own: riders’ accommodation Riders’ length of employment as courier riders 5. DRIVER ORGANISATION: TA W C h T ic IC h S c , o C m H p A an LL y E d N o G y E o S u A rid N e D fo S r U ? CCESSES Structural ch P a A lle R n T g 2 e : s TtoHEorNgAaTnUisRinEgAinNDtheWeO-hRaKiIlNinGg sCeOctNoDr ITIONS OF A BIKER’S JOB Strategies used in driv1e.r Roirdgearns’ispinegrception of employment relationship: a worker or self-employed partner? Key milestones: attem2p.tsWahnadt sautctrcaecstseeds riders to working as food couriers? Attempts to 3st.rWikeasbbyedinrigvearsfood courier riders’ only occupation? Compromise4s.mRiaddeersbwy hdorivoewrsniendstvrsi.kereanctteidontheir motorbikes Hurdles ahead: critical5s.tIesptshe nature of the relationship between the riders, the courier company and Pressing chaltlheengceusstomer neutral or biased? Addressing t6h.eIsnetheed rfiodrergoravteinrngmsyesntteaml siumpppoorrttant to you? 7. The link between the rating system and riders’ commission Addressing outstanding issues from UBER Conclusion 8. Hours of work and unpaid labour 9. Problems couriers faced at work PART 3: 6. TURNING THE CORNER: FORGING A WAY FORWARD WHAT POWER DO FOOD COURIERS HAVE TO CHALLENGE THE UNEQUAL Regulatory reform WORK RELATIONSHIP? Worker cooperatives Extract from Fikile Masikane’s field notes journal Wo 5 r . ke B r U c I e L n D t I r N es G WORKER ORGANISATION FROM BELOW Online forumTshe emergence of union-like associations Conclusion Organising digital platform workers in Kenya The Power Resources Approach(PRA) 7. REFERENC Es E s S ential capabilities to be developed 6. LOOKING TO THE FUTURE 7. ACKNOWLEDGEMENTS References 6 3 2 7 7 8 10 11 12 14 15 16 16 17 19 19 20 20 20 21 21 22 22 22 23 23 25 25 26 27 27 30 31 32 32 33 36 38 42 43 43 44 46 49 51 51

PRELUDE: A DAY IN THE WORKING LIFE OF TWO FOOD COURIER RIDERS A dangerous job: “It’s a Tuesday morning. I’ve finally decided to take an off day. I wake up at 11am. I do my usual routine of going onto WhatsApp. I notice a familiar face on many statuses. It’s Abdul. We call him‘Ntshebe’ because of his long beard. His photo accompanied by the captions‘RIP’ or‘gone too soon’. I’m totally shocked. I just saw him yesterday. I call one of my friends to confirm. Indeed‘Ntshebe’ is gone. He’s Muslim, meaning he’s going to be buried on the same day of his death. I gather more information and I’m told his burial time. I make arrangements to get there. There’s more than 100 motorbikes. We accompany the hearse to Albertville Cemetery where‘Ntshebe’ is laid to rest. Abdul was one of many hit and run victims in our line of work.” – Brian Chibwe 1 Journey to Johannesburg : Joseph Apio 2 is a 38-year-old rider from Uganda. He is a self-acclaimed activist and chairperson of the association, United Ugandans(UU) and Brothers of Melville(BOM) Johannesburg in South Africa. He comes from a poor background; he studied in a small village in Uganda. Joseph could only go up to O-level because of financial constraints. 3 He has eight younger siblings who he must take care of, in addition to his wife and four children. Hopeful, Joseph says,“This right now is my chance. I came to South Africa for better opportunities. When you are broke, you do whatever it takes to get out of that situation. I had to convince my wife that I must leave. I told her I had to leave and go to SA[South Africa]. We saved money, and I came here.” Determined to start this journey to seek a better life with better opportunities in South Africa, Joseph travelled by bus. It took one week to reach South Africa. Arriving in Johannesburg was a terrible experience for him. Thieves robbed him at Wanderers Taxi Rank where passengers got off the bus. Narrating the incident, Joseph says,“That was my first experience of SA, but I had to be firm to survive. Since I grew up in a tough situation, I had to survive. I always feel like I do not belong here.” – Joseph Apio 2 1 A food courier’s two-week diary entry, August 2020. 2 Not courier rider’s real name. 3 Interview with Joseph Apio, food courier motorbike rider(8 July 2020). 7

THE RISE OF PLATFORM CAPITALISM “The fetishism of the wage may well be the source of capitalist ideologies of freedom and equality, but the employment contract is not the founding moment. For capitalism begins not with the offer of work, but with the imperative to earn a living.” – Michael Denning,“Wage-less life”. New Left Review 66(November/December, 2010:80). A new worker has emerged in the digital economy; this new worker is subject to a new business model based on a form of authoritarian algorithmic management. This management is controlled by a few corporations, popularly known as tech giants, such as Uber and Amazon. These companies possess unprecedented levels of power, ostensibly based on‘freedom’ and‘selfemployment’. These companies are deepening worker insecurity and undermining worker rights. They are dramatically increasing inequality between a core group of extremely wealthy senior company managers and a growing pool of precarious workers. This report seeks to answer three critical questions: Who are these new digital workers? What are their working conditions? How can these workers improve their working conditions? Many countries have embraced the promise of the digital economy with enthusiasm(Webster, 2020). The Taylor Report(2017), which was based in the United Kingdom (UK), argued that platform or gig-based work provides welcome opportunities for those who may not be able to work in more conventional ways(Taylor Report, 2017). Similarly, in South Africa, an initiative with presidential support – South Africa in the Digital Age(SADA) – sees significant income opportunities in three broad areas: globally-traded services, labour-absorbing platforms, and technology hubs(SADA, 2019). For many young people desperate for an income, it seems like platform or gig-based work can provide them with an opportunity to have a job and be‘selfemployed’. Indeed, because of the apparent‘flexibility’ and‘freedom’ that comes with being a food courier rider, many believe they are in‘partnership’ with Uber Eats or Mr Delivery, as opposed to simply being employed by such companies. This platform business model has transformed the employment relationship and undermined worker rights. The model creates a highly segmented labour market comprising a small core of high value-added activities, and a non-core of outsourced and franchised activities. On the one hand, the core workers enjoy enhanced salaries, pensions and other benefits. Their founders and CEOs, such as Jeff Bezos, Bill Gates and Mark Zuckerberg are among the richest business owners in the world. In late June 2020, Jeff Bezos, for example, added US$13 billion to his net worth in a single day, illustrating the extraordinary wealth of such tech giants. 4 On the other hand, the workers at the peripheral or outsourced outlets, must make do with far inferior and often precarious pay and working conditions. We refer to this as the‘Uberisation’ of work (Webster, 2020). Moreover, the choice facing food couriers can be compared to the dilemmas described in Elliott’s(2013)“Suffering Agency”, where he explores the painful choices vulnerable people face in the context of an unequal neoliberal world. In a nutshell, these new business entities concentrate on high value-adding activities while divesting themselves from‘downstream’ employment liabilities. They do this through technology-enabled outsourcing and subcontracting practices that manage their fragmented supply chains remotely. 8

The tech giants’ practices rest on three key characteristics: 1. They display‘monopoly tendencies’ as exemplified by the trillion US$ valuations many of them have achieved, and their drive to undercut other producers (Srnicek, 2016, p. 48; see also Nocke et al., 2007). 2. They are willing to by-pass standard corporate governance norms, and have a particular appetite for dual-class shares which deliver their founders huge salaries and very extensive share options (Govindarajan& Srivastava, 2018). 3. In their employment polices,‘riders’ are misclassified as‘partners’ and designated as‘independent contractors’ with a‘self-employed’ status. Rider ‘partners’ provide the tools of their trade and are paid on a piece work basis – but not for their working time. Due to their so-called‘self-employed’ status, the company by-passes workers’ rights covered by standard employment relationships. Platform capitalism has created a new work paradigm where workers are managed through online platforms, monitored indirectly, and expected to produce measurable outputs(Huws, 2016). Work is‘logged in’, Huws explains, in three distinct ways: it is divided into standard and quantified components; it is subjected to continuous surveillance and monitoring; and it requires a worker to be connected to an online platform to obtain work(ibid., p.15). Instead of clocking-in as at a traditional workplace with their timecard,‘gig’ workers log into an‘app’. In so doing, they become subject to a new business model based on a form of authoritarian algorithmic management that: • Translates consumers’ demands into orders workers must deliver • Determines what tasks workers must execute, where and when • Directly or indirectly determines how much money workers will be paid for the execution of particular tasks • Through the algorithm, directly or indirectly controls the execution of the work and the worker’s performance at work. Drawing on labour process theory, it is possible to identify three common features of digital work. Firstly, there is a common point of production, the app, which is downloaded on a smartphone. The app is where workers and customers encounter each other, and it is the app that operates as the point of production. The platform itself determines the worker’s percentage of what the customer pays. This percentage is based on algorithmic elaborations, the specifics of which are blocked from workers. The worker, in this instance, the rider, either accepts or declines orders. They cannot intervene in the calculation. If they do, the platform can deactivate the rider’s app. Secondly, embedded in the worker’s execution is customer feedback – customers rank and rate the riders. The platform translates this into a reputational score, a proxy of trustworthiness. Hochschild(1983), in her classic study of flight attendants in the United States (US), describes this as a form of emotional labour.“This labour,” she says,“requires one to induce or suppress feelings to sustain the outward countenance that produces the proper state of mind in others – in the case of flight attendants, the sense of being cared for in a convivial and safe space”(ibid., p. 7). Thirdly, a new form of managerial control is introduced that Gandani(2018) calls a form of“...technonormative control over workers:(a) platforms are akin to management by customers, like call centres;(b) Personal Bests(PBs) are used to stimulate and reward workers. These forms of control make invisible the management figure; they become hidden and inaccessible”(Gandani, 2018, p. 15). 8 Gig work exemplifies new market-based principles where precarious employment relations, along with algorithmic controls of the labour process, are used to great effect in shifting risks from capital to labour. Furthermore, 4 https://www.forbes.com/sites/arielshapiro/2020/10/10/jeff-bezos-and-ex-wifemackenzie-scott-gain-12-billion-in-anticipation- of-amazon-prime-day-trumpcovid/#122ca6cb1df2[3 July 2020]. 5 https://static1.squarespace.com/static/588c7bb829687f7716ffe2ea/t/5892e76 4ebbd1a9a7ffb919f/1486022526732/ Elliott+Suffering.pdf[18 November 2021]. 6 Some classes of shares give‘executive owners’ the power to control a corporation’s decision-making without having a voting majority. This they do in exchange for taking a lower share of profit. 7 “Software application used by the platform”, see: https://www.ncbi.nlm.nih. gov/pmc/articles/PMC4129112/[29 November 2021]. 8 Interestingly, the executive owners of giant tech companies are often very visible as philanthropists, contributing substantially to worthy causes, but largely invisible when it comes to questions of employment and workers’ conditions of work. 9

gig work favours individual freedom over collective freedom, which puts further pressure on workers’ ability to control their wages and working hours. As a result, the alleged‘freedom’ in gig work, which offers workers flexibility to schedule their working and personal lives freely, is heavily constrained(ibid., 2018; see also Anwar & Graham, 2020, p. 16). We interviewed a senior executive of a major tech company to understand his perspective on the debate as to whether the riders were independent contractors or employees. This was his reply:“Hiring these individuals in the independent labour market is at least three times the cost of that labour and it then increases, and the total space of the market decreases and therefore the economic opportunity decreases. We surveyed the drivers and what we found is that drivers like having their own micro-enterprise. The polls say+70% said [the riders saw themselves as] independent contractors; this works for our business. Contractor Law in California also agrees with that. We have a model which provides health and sickness benefits, so they can claim; there is a lot involved there. But it’s a model we would like to see go around the world. From a policy point of view, it allows us to have a bigger supply and more participants, where people have economic opportunities. It is a much better plan for economic opportunities”(Zoom interview, 13 November 2020). However, the Indian Federation of App-based Transport Workers(IFAT) highlight the health issues and/ or ailments that app-based workers experience due to their work, and also looking at the harassment they face while performing their work(IFAT, 2020). The food courier riders in this study highlight how they have either experienced hijacking, robbery, or theft, or know someone who has. Other riders have been involved in road accidents. They know of colleagues who have been killed whilst executing their work. AUTHORITARIAN MANAGEMENT? Cant(2020:59) calls the app, authoritarian management:“The App spits out a sequence of repetitive commands and you just have to do it.” The app makes all the decisions because it has all the information (ibid., p. 61). The app deskills workers because they just have to follow instructions from programmes created Diagram 1: Tassinari and Maccarrone’s(2020) four-sided model of food delivery platforms Source: Tassinari and Maccarrone(2020) by software engineers(ibid., p. 62). ln light of the individualisation, dispersal and pervasive monitoring that characterise work in the‘gig economy’, Tassinari and Maccarrone(2020) argue that the development of solidarity among gig workers seemed unlikely. However, numerous recent episodes of gig workers’ mobilisation required them to reconsider this assumption. In the UK and Italy, workers demonstrated the processes through which workplace solidarity among gig workers could develop, in two cases of mobilisation of food delivery platform couriers. Through the framework of labour process theory, Tassinari and Maccarrone identified the sources of gig workers’ antagonism in the app-mediated model of work organisation, and the factors that facilitate and hinder the consolidation of active solidarity towards the emergence of collective action. Their article emphasises the centrality of workers’ agency in overcoming constraints to solidarity and collective action, and the diversity of forms through which solidarity can be expressed in hostile work contexts. Two central factors enabled solidarity between couriers in the UK and Italy cases: the availability of free spaces; and the existence and nurturing of social relations. Beyond the initial collective expression of grievances, such as police harassment and accidents, respondents in the study highlighted that a sense of shared identity as gig 10

Photograph 2:“We have to wear protective gear. There are many accidents in this line of work.€“ Riders’ Photo Competition, Hillbrow, Johannesburg, South Africa(September 2021) by Justice Tafamba. workers was forged through their very lived experience of mobilising together. The experience of protesting in large numbers outside the different companies’ offices and in the streets, empowered the protesting workers. It led to couriers realising that taking action was possible and, in some respects, also easy(ibid., 2020). What enhances digital workers’ bargaining power? Cant(2020) in his account of global food delivery company Deliveroo, argues that digital management methods facilitate courier mobilisation because the method offers a space for selforganised courier associations to boost their associational power(ibid., p. 16). He debunks the argument around social entrepreneurship, arguing for the need to translate workers’ gains into rulemaking and institutional power, such as the demand for a national minimum wage. Vandaele(2018), from the European Trade Union Institute(ETUI), points to the potential of discursive or social power, such as coalitions and the media being “translated into rulemaking and institutional power”. This will allow the State to take on“responsibility for regulating employment relations in the platform economy, by setting minimum standards on wages and social protection”(Vandaele, 2018, p. 16). We address the potential of such social power in the Power Resources Approach, which we present towards the end of this report. Cant(2020) further argues that certain forms of collective representation of platform workers is emerging(ibid., p. 18), including grassroots unions, union-affiliated guilds, traditional unions, labour market intermediaries, such as labour mutual or quasi-unions, and worker-led cooperatives. Importantly, platform workers are now attracting the attention of organised labour. For example, in 2018, the Transport Workers Union of Australia protested against Deliveroo’s policy of categorising food delivery workers as independent contractors instead of employees. Partnering with the Show Drivers Cooperative and Delivery Riders Alliance, they have campaigned for a minimum wage and the right to bargain. Workers successfully forced global food delivery company, Fodora, to pay back unpaid wages(Chan, 2019). A DOUBLE-EDGED SWORD The new digital technology, we are suggesting, is a double-edged sword: on the one hand, it is leading to an extension of authoritarian managerial control over workers, increasing their insecurity and deepening levels of inequality between workers and senior management. On the other hand, by technologically linking workers, management has increased workers’ bargaining power, thereby providing workers with the ability to develop and mobilise collective solidarity, including strike action. The unprecedented wealth in the hands of company directors is generating resistance to the precarious employment conditions in these tech giant companies. Courier riders’ actions are leading to increasing attempts to regulate the sector. In California, for example, legislators approved a landmark bill that required companies such as Uber and Lyft to treat contract workers as employees(Conger& Scheiber, 2019). 9 https://www.latimes.com/california/newsletter/2021-08-23/proposition-22lyft-uber-decision-essential-california. 11

This bill was, however, overturned by the courts in November 2020, and food couriers’ status as independent contractors confirmed. 9 There have been similar attempts by transport workers to get Uber taxi drivers declared employees in South Africa. 10 Hybrid forms of union-like organisations are emerging side-by-side with traditional trade unions to defend workers’ needs and interests in the digital economy. Historically, the world of work has constantly been transformed through what Silver (2003) calls‘technological fixes’. These technological fixes lead to new types of workers, new forms of worker organisation, and new modes of struggle. Despite the obstacles the platform model of work organisation poses, worker solidarity in the gig economy is possible, and is rooted in the structural antagonism intrinsic in the labour process. Its development, however, is not mechanistic; it is shaped by workers’ agency and by a diversity of contextual factors(Tassinari& Maccarrone, 2020). AFRICA’S PLATFORM ECONOMY The rise of the gig economy in Africa is a direct result of greater access to the internet. It is estimated that between the years 2000 and 2020, Africa experienced a 12 153% increase in the number of internet users, with an estimated 600 million Africans now online. This, however, is still less than half the population of Africa(Internet World Stats, 2021). Similarly, research into digital platforms by insight2impact showed that the number of digital platforms increased by 37% between 2018 and 2019. In 2018, these platforms had already been estimated to provide income-generating opportunities to 4.8 million workers across seven countries in Africa (Johnson, Dunn and van Vuuren, 2020). South Africa has the most digital platforms, overtaking Nigeria in 2018; while Kenya has experienced the fastest growth in platforms. Thus, the relentless drive to on-board more Africans continues, as indicated in Diagram 2. There are three aspects to accessing the digital economy. We focus on the first two in this report: 1. Digital infrastructure(cables) 2. Access to digital devices(tablets, cellphones) 3. Digital knowledge(how to use the device, how to accept/ make/ receive payments; how to accept work). Diagram 2: Africa’s digital platforms 11 Digital infrastructure It is often forgotten that the gig economy requires extensive foundational infrastructure, including massive investment in subsea cables. The World Bank estimated that the infrastructure investment needed in Africa to achieve universal access is US$100 billion(World Bank, 2019). This investment cannot be handled by one entity; it requires many different actors in the private and public sectors to achieve the broadband investment needed. As an example, infrastructure investment in Africa has also attracted the attention of the tech giants. Facebook, partnering with telco companies in Africa, launched an audacious project dubbed 2Africa, aimed at laying approximately 37 000km of underwater fibre cable along the coast of Africa. 10 http://www.saflii.org/za/cases/ZALCCT/2018/1.html[10 June 2021]. 11 Johnson, Bester, van Vuuren& Dunn.(2020).[image] Africa’s digital platforms: Overview of emerging trends in the market. The Centre for Financial Regulation& Inclusion. See: https://www.researchgate.net/publication/341940397_Africa%27s_digital_platforms_overview_of_emerging_ trends_in_the_ market. 12

13

Not to be outdone, Google launched a similar project Equiano, aiming to construct a cable between Lisbon in Portugal and Cape Town in South Africa, with a further nine branching units along the way. Diagram 3 shows the high concentration of trans-Atlantic and trans-Pacific subsea cables connecting North America to Europe and North America to Southeast Asia. Unity(US, JP) 2010 FASTER(US, JP, TW) 2010 PLCN(HK, US) 2019 Curie(CL, US) 2019 Havfrue(US, IE, DK) 2019 Dunant(US, FR) 2020 Equiano(PT, NG, ZA) 2021 Monet(US, BR) 2017 Diagram 3: Subsea cables across the world providing access to digital devices 12 Junior(Rio, Santos) 2018 Tannat(BR, UY, AR) 2018 Indigo(SG, ID, AU) 2019 SJC(JP, HK, SG) 2013 HK-G(HK, GU) 2020 JGA-S(GU, AU) 2019 Access to digital devices As Graph 1 illustrates, access to mobile technologies is highly unequal across Africa. For instance, in South Africa, more than 80% of the population has access to a cellphone. In Mozambique, this figure is just over one-third. Access to cellphones is highest among men in urban centres and lowest among women in rural areas, though the degree of inequality varies. In South Africa, for example, men are 4% more likely to have cellphones than women, whereas in Mozambique the gender gap is as wide as 14%. Disparities are even wider when one looks at internet usage. In South Africa, 56% of the population has access to the internet. However, if one looks at coverage at the household level, only 11% of households have access to the internet. Although South Africa has higher levels of internet coverage than most countries on the African continent, access to the internet is concentrated in a minority of households; this is, however, expected to grow. Next, we look at the history of food courier companies entering, and occassionally, initiated in Africa. 100 90 80 70 60 50 40 30 20 10 0 Ghana Mozambique Kenya South Africa Colombia India Ethiopia Access to mobile phone Access to internet Access to mobile banking Graph 1: Access to mobile technologies Source: Castel-Branco, Mapukata and Webster(2020) FOOD COURIERS ENTER AFRICA Uber began operating in South Africa in 2013; it launched Uber Eats in 2016. 13 What is now known as Mr Delivery was founded in Cape Town in 1992. It was fully acquired by Takealot(an e-commerce company) in 2014. In 2015, the company began to shift from a phone call-based ordering process to an app-based ordering 12 https://techcrunch.com/wp-content/uploads/2019/06/Cloud-Map-with-Equiano-FINAL.png. 13 https://www.businessofapps.com/data/uber-eats-statistics/[14 November 2021]. 14

process. In June 2016, the business consisted of 80% phone calls, and by June 2017 it had become 95% appbased. 14 In 2015, a merger between Takealot and the Naspers-owned firm, Kalahari, was completed. In 2017, Naspers increased its investment in Takealot by 96%. By implication, Naspers, a South African firm, now owns Mr Delivery. Kute, Molife and Tshezi(2020) give two reasons for the rapid entry of Uber into Africa. Firstly, relaxed regulations in Africa significantly lowered Uber’s barriers to entry, thereby enabling its rapid expansion and job creation. Secondly, Uber made shrewd and agile business model adjustments to adapt to Africa’s unique operating environment. This aided its continued growth and expansion on the continent. Globally, online food couriers are operating at a net loss, as they attempt to increase their market share by spending large portions of their revenue on advertising, promotions, bonuses and exclusive partnerships with big brands such as MacDonald’s. MacDonald’s makes up just under 9% of all outlets on Uber Eats globally. 15 Further attempts to gain market share have been initiated through setting very low delivery prices: 45% of revenue goes towards driver incentives and driver referral payments. 16 Uber, including Uber Eats, made an US$8.6 billion loss in 2019. However, gross bookings grew by over 15 million in 2020 to US$65 million. According to an investor presentation, losses are smaller each year and there were plans to break even by 2021. 17 The percentage of food delivered from restaurants to customers in South Africa was 16.1% in 2020. Projections are that it will reach 20.4% by 2024. 18 As of 2018, Mr Delivery saw an order growth of 210%. 19 FOOD COURIER COMPANY REVENUE Uber Eats and most food couriers are only profitable in one quarter of the markets in which they operate. 20 However, lower losses have been reported by online food platforms each year. 21 Uber’s earning release indicates that globally during the COVID-19 pandemic, ride hailing significantly decreased, but food couriering dramatically increased. However, although there were more Uber Eats customers than Uber ride hailing customers, the taxi drivers are still the dominant contributor to the company. Uber Eats’ global revenue was US$600 million in 2017, and increased to US$1.2 billion by the second quarter of 2020. 22 There were 13.5 million online food delivery customers during 2020. This represented a 27.5% increase in growth from 2019. The online food industry generated a revenue of US$965 million, representing a 35.4% increase in growth from 2019. South Africa’s level of food courier growth is higher than most other countries. 23 Globally, restaurant-to-consumer delivery accounts for 9.6 million users and US$442 million revenue. 24 Mr Delivery has an estimated 2 million downloads and 700 000 active monthly users. Uber Eats has 2.1 million downloads, but in 2020 did not make public its sales data(2020). 25 Naspers CEO, Larry Illg, indicated that food delivery is still an under-penetrated market in South Africa and Mr Delivery was still attempting to expand. 26 14 https://www.itweb.co.za/content/2j5alr7QxBDvpYQk[18 June 2020]. 15 https://www.theverge.com/2018/3/12/17110008/uber-eats-food-delivery-expanding[12 November 2021]. 16 https://www.restaurantdive.com/news/breaking-uber-eats-lead-steps-down-amid-profitability-struggles/572976[8 June 2020]. 17 Investor Presentation_2020_Q4(q4cdn.com)(23 September 2020). See: https://s23.q4cdn.com/407969754/files/doc_financials/2019/sr/InvestorPresentation_2020_ Feb6.pdf. 18 Uber Eats Revenue and Usage Statistics(2021) – Business of Apps(8 August 2020). See: https://www.businessofapps.com/data/uber-eats-statistics/. 19 https://www.businessinsider.co.za/fast-food-deliveries-are-exploding-in-south-africa-here-are-the-winners-2018-8[23 September 2021]. 20 https://www.theverge.com/2018/3/12/17110008/uber-eats-food-delivery-expanding[23 September 2021]. 21 User penetration in the Online Food Delivery segment is estimated to be 23.1% in 2021. Online Food Delivery – South Africa| Statista Market Forecast(29 November 2021). See: https://www.statista.com/outlook/dmo/eservices/online-food-delivery/south- africa?currency=usd. 22 https://www.businessofapps.com/data/uber-eats-statistics/[16 July 2021]. 23 Restaurant-to-consumer: User penetration is estimated to be 21.0% in 2021 and is expected to hit 30.8% by 2025. Restaurant-to-Consumer Delivery – South Africa| Statista Market Forecast(29 November 2021). See: https://www.statista.com/outlook/dmo/ eservices/online-food-delivery/restaurant-to-consumer-delivery/south-africa. 24 https://www.statista.com/outlook/375/112/restaurant-to-consumer-delivery/south-africa[16 September 2021]. 25 The online food industry generated a revenue of US$717, representing a 27.2% increase in growth from 2019. See: https://www. statista.com/outlook/dmo/eservices/ online-food-delivery/south-africa?currency=usd. 26 https://www.businessinsider.co.za/fast-food-deliveries-are-exploding-in-south-africa-here-are-the-winners-2018-8[8 June 2020]. 15

RESEARCH METHODS This section explores the various methods we employed to conduct the study. It highlights the opportunities and challenges while conducting research during the COVID-19 pandemic. However, despite the pandemic and digitalisation, ethnographic research remains the best method to capture the experiences of participants. THE CHALLENGE OF CONDUCTING ETHNOGRAPHIC RESEARCH DURING THE COVID-19 PANDEMIC When our study began in February 2020, our main intention was to conduct a survey across three cities. We planned to spend time in the field understanding the lived experiences of riders in Accra(Ghana), Johannesburg (South Africa), and Nairobi(Kenya). Our intention was to employ a horizontal mapping approach where the workers themselves would be trained to conduct the research on their work and working conditions(Webster & Bischoff, 2011). However, the COVID-19 pandemic and subsequent lockdowns prevented us from travelling to these countries to train the researchers, and we ended up employing researchers in the respective countries. In early March 2020, we conducted a pilot study in Campus Square, Auckland Park, Johannesburg. We spoke with two riders who gave us an idea of their daily lives and experiences of being riders in Johannesburg. We spent some time observing the nature of their work, how long it took for them to get orders, and how they interacted with each other as they waited for their order requests. Shortly after this, the South African Government instituted a very strict Alert Level 5 Lockdown due to the COVID-19 pandemic. This meant that no one was allowed to go out of their homes unless it was for essential shopping or with a special permit. This affected people in various ways; it particularly affected the food courier riders where there is a policy of‘no work no pay’. This brought our study to a standstill. We could not interview the riders as planned. Months went by, and as things started opening slowly, for riders’ safety and our own, we decided to conduct the survey interviews online using WhatsApp video calls across all three cities. As submitted and stated in our ethics clearance, we offered riders a small compensation of US$13: R200 in Johannesburg; KSh1,486.32 in Nairobi; and GH₵81.20 in Accra, for data and time with us that they would otherwise have spent working. We knew that the nature of their work was precarious and informal, and that riders make money by the hour. We collected all riders’ cellphone numbers and set out the times for interviews. While we recognise that conducting online research is not impossible, it was, however, suboptimal particularly for this study. We were unable to travel to Accra and Nairobi to go through the questionnaires face-to-face. Instead, we set up weekly meetings with the field researchers to see how the surveys were running. One of the major challenges we faced was how the researchers interpreted and understood the questions. Although we had the same survey questions in all three cities, some of the questions for Accra and Nairobi had to be rephrased to help the researchers explain them to the riders. Being unable to travel and engage in face-to-face training was a major impediment created by the COVID-19 pandemic. We were not able to extend our research strategy into action research in Accra and Nairobi as we did in Johannesburg. However, despite the challenges presented by the pandemic, we believe that by extending our research strategy to include several additional 16

instruments, we have been able to satisfactorily capture the lived experiences of food courier riders. RESEARCH STRATEGY We drew on several qualitative research methods. The first involved in-depth interviews in Johannesburg and one in-depth interview with a woman rider in Nairobi. In addition, we constructed biographies of two riders in Johannesburg. We collected a diary entry from one rider who wrote of his experiences via WhatsApp over a period of two weeks. We also participated in a WhatsApp group called‘Unionise food delivery riders in Johannesburg’. This group had 13 food delivery riders. They used their WhatsApp group as a platform to write about their experiences and share information about work. This gave us deeper insight into what it means to be a food courier rider in Johannesburg. It helped us understand Uber Eats and Mr Delivery’s business models. We also conducted an in-depth interview with Lawrence Dlomo, who works in the e-hailing business. Our main research instrument was a semi-structured questionnaire to interview 150 riders drawn equally from Accra in Ghana, 27 Johannesburg in South Africa, and Nairobi in Kenya. 28 We interviewed riders in and around shopping centres and malls where there was a high concentration of food courier riders. In Accra, riders ride for Jumia Foods, Bolt Foods, Glovo, Solar Taxi (which seems to operate as a third party), 29 and Uber Eats. We used a snowball sampling technique to select 50 participants in each city. The survey questions were prepared in English and took between 30 and 45 minutes to complete. Interestingly, riders in Johannesburg seemed to be divided by nationality into different geographical areas. For example, riders at Campus Square, Columbine Square and Cresta were originally from Uganda. At Rosebank Mall, most riders were originally from Zimbabwe, with a few from the Democratic Republic of Congo(DRC); at Clearwater Mall most of the riders were originally from Zimbabwe. The riders in Johannesburg were largely undocumented cross-border migrants who did not have work permits, or whose permits had expired. Reasons for riders’ reluctance to participate in our study included the threat of losing their jobs because some do not have the correct documents to be in South Africa; and their fear of who we represent, and why we were conducting the study. For that reason, anonymity was always ensured by not writing down riders’ names on the questionnaires. In all three cities, where consent was not given, riders’ real names are not used in this report, to protect the participants from being identified. We engaged in participant observation by attending soccer tournaments the riders hosted on public holidays in South Africa. This led us to visual ethnography. We introduced a photo competition among the Johannesburg riders. Participants submitted 10 photographs each that illustrated their experiences as riders. They included short narratives to describe what it meant for them to be a food courier riders. Some of these photographs are included in this report. To offer us greater insight into platform work, in June 2021 co-author of this report, Fikile Masikane, registered as an Uber Eats rider. 30 We turn to this later in the report. In the process of reporting back on our findings to our respondents, an on-going forum emerged. In essence, our research became a form of critical engagement or participatory action research. Participants sought transformative change through the simultaneous process of acting and doing research, linked together by critical reflection(Lewin, 1946; Lozano, 2018; Cossyleon & Spitz, 2021). 31 The purpose of the forum was, and continues to be, to create a structure where riders can air their grievances across the different work zones, and collectively find ways to address grievances with the relevant stakeholders. 27 Due to COVID-19 and the lockdown restrictions, we arranged field researchers to assist with the surveys in Accra and Nairobi. Karim Saagbuk was our field researcher in Accra, Ghana. 28 Agnes Tsheri was our field researcher in Nairobi, Kenya. 29 Here, riders said they were‘employed’ by Solar Taxi as a third party, instead of stating, like other riders, that they rode for Uber Eats or Jumia. 30 See“Registering as a food courier driver” under Research Findings. 31 Critical engagement addresses the difficult challenge of producing knowledge in a collaborative way while advancing research that might be meaningful for both the social sciences and the constituencies with whom the research is undertaken(Lozano, 2018, p. 107). Instead of working on social movements, Lozano suggests a collaborative approach which involves “working and thinking together with social movements’ activists, advancing research that might be meaningful for both social science and the‘research subjects’”(Lozano, 2018, p. 103). 17

18

SEEING THE RIDERS’ WORLD OF WORK THROUGH THEIR OWN LENSES Under the Future of Work project, SCIS invited Johannesburg food courier riders to enter a Riders’ Photo Competition, organised between August and September 2021. Action research: the seeds of a rider organisation In March 2021, we invited the riders we had interviewed in Johannesburg in 2020 to a forum on the Wits University campus. The initial purpose of the forum was to report back on our findings and to invite the couriers to respond. Participants endorsed our key finding that riders become food couriers to survive, not because it gave them greater freedom and flexibility. Riders seemed to be saying that any job is better than no job at all. However, in discussing their working conditions, riders saw the need to formally organise in order to challenge their‘employer’, but they felt they had no power. This is of course not true, as they had already exercised their power in December 2020 by going on strike. We invited international informal-worker organiser Pat Horn to facilitate a discussion on riders’ challenges. She argued that the‘self-employed’ are workers and can be organised into unions, as they have been in the SelfEmployed Women’s Association(SEWA) in India. Riders would not be taken seriously, said Horn, until they organised collectively, had representatives to put their demands to the companies, and engaged in collective bargaining. A rich discussion followed. Riders raised grievances, particularly that of not having occupational health protection. For example, in the event of a robbery or accident, the company does not compensate the riders. One of the participants, in fact, was involved in an accident. An ambulance came to pick him up, and he was later charged for the cost of the ambulance, which he did not even call.* As a response from the first forum, the riders suggested an on-going forum to assist them in being more formally organised. Fortunately, an embryonic organisation already existed through WhatsApp groups. The challenge was to strengthen the groups and provide a catalyst for building a sustainable, permanent member organisation from below. In the next forum, Horn and trade union educator, Dinga Sikwebu, presented the history of the formation of trade unions in South Africa, and especially the role of cross-border migrants in forming these unions. Horn suggested how to establish a movement, and how to go ahead and challenge their employers. Riders discussed practical ways to do this, particularly as cross-border migrants. Fifteen riders participated in the third forum. Here, co-author, Fikile Masikane, presented the findings of our research in Accra and Nairobi, emphasising the similarities and differences of the couriers’ experiences across the cities. The Johannesburg riders established an interim committee as a step towards organising a bigger and more formalised democratic association. Since then, there has been one more meeting with the interim committee, with Fikile Masikane as one of the convenors. Riders decided to demand that the company allocate riders a physical office they could go to when faced with work challenges. In this meeting, a possible name of a worker organisation was chosen by the riders: the South African Food Couriers Association(SAFCA). September 2021 * Uber Eats GM for Sub-Saharan Africa, Nkampe Molewa, claimed“... all [its] delivery people have injury protection from AIG Insurance...”[Katherine Child. Uber Eats sees uber shift in behaviour as here to stay. Business Day, 26 November 2021 p. 13], something the drivers deny knowledge of. 19

RESEARCH FINDINGS PART 1: WHO ARE THE FOOD COURIER RIDERS? A DEMOGRAPHIC PROFILE In this section we discuss the demographic profile of the riders. They have been doing the job for some time(over three years) across the three cities, and predominantly live in rented accommodation. There are important national variations: in Johannesburg, 90% are foreign nationals, and in Nairobi and Accra riders are all local citizens. The surprising finding is that the workforce is relatively stable; riders remain in the job for relatively long periods of time. The most likely explanation is that app-based platform work is the best job available to them. As respondents remarked,“I just want to survive.” However, our Nairobi data suggests that for those 12% of courier riders who have been in the job for over five years, food couriering may well be a career. Further on in the report, our graphs reflect our findings in more detail. Here, to highlight life as a rider, we begin with co-author, Fikile Masikane’s account of registering as a food courier driver. REGISTERING AS A FOOD COURIER RIDER IN SOUTH AFRICA, by Fikile Masikane After spending several months interacting with riders, I decided to join the app as a rider myself, to get a deeper sense of riders’ context in South Africa. In May 2021, to riders’ surprise, I asked how to register as an Uber Eats rider. They said it was simple and straightforward. I first downloaded the Uber Eats app on my phone and went to the tab:‘Deliver with Uber – partner with Uber and earn money in your spare time’. The app provided information on the standard requirements: a motor vehicle or a motorbike, a South African driver’s licence and ID photograph, ID number/passport, a work permit, asylum letter(if not from SA), and a police clearance. Photograph 3: ID photograph Masikane submitted with other required documents to Uber Eats(May 2021) by Fikile Masikane. Simple. I thought so, too. I uploaded the documents and it seemed easier for me since I have a South African ID and driver’s licence. The process began on the 7 May, 2021 by going to Tiger Wheel and Tyre, a private company in Fordsburg, Johannesburg to get my vehicle inspection done. This cost R180(US$11.50). The first question the assistant asked was:“Where are you from?” I replied,“Parkhurst.” The inspector said,“No, your nationality?”“South African,” I replied. With what seemed like a smile, he said,“Great, it is straightforward. Give me your ID book and we will begin.” I spent about an hour waiting for my turn and for the inspector to do the inspection.The following day I had to get a police clearance check. 32 A rider told me it was done at the Methodist Church in Parktown. This was interesting to note, as one would expect to get the clearance at a police station. 32 From Uber Eats website[27 November 2021]: “Partners are required to pass a criminal background check through National Crime Check(NCC). This can be done online and is usually processed within 14 days. You will receive a confirmation email and a copy of your results once your background check has been processed. No action is required from you after lodging – NCC will supply a copy of your results to Uber.” 20

The clearance cost R500(US$30.37). The assistant at the church was an Uber Eats employee. He took my fingerprints and sent them to the South African Police Service(SAPS) headquarters in Pretoria on the same day. I spent about an hour to get these done. On 9 May 2021, I received an email from Uber Eats to notify me that SAPS had sent a Personal Credential Verification Report. Soon after, I emailed all my documentation to Uber Eats. A week later, I received feedback that my vehicle inspection failed on one window wiper. The inspector had not reported this to me when I went there with my vehicle. I then had to return to Tiger Wheel and Tyre to redo the vehicle test, luckily at no cost. When all my required documents were successfully completed, I submitted them to Uber Eats. They told me to pay an additional fee of R1,000(US$69.98) to receive the Uber Eats branded delivery bag and jacket to secure my place. Uber Eats further communicated that when all documentation and proof of payment was received they would allocate me a slot to start the delivery service. It has been over six months since I was registered, and by June 2021 I had not heard anything from them. This all cost me R1,580(US$110.56). Nationality of riders The most striking difference in our three-country sample of riders was the high percentage(94%) of cross-border migrants in Johannesburg. Riders come from Uganda, Zimbabwe, Malawi and the Democratic Republic of Congo(DRC). The result is a multiplicity of languages among riders. Johannesburg has been the magnet for migrant workers since the nineteenth century, but in post-apartheid South Africa there has been a dramatic increase in cross-border migrants from neighbouring countries. In contrast, all the riders in Accra and Nairobi were local nationals. Johannesburg riders are more precarious and vulnerable than their African counterparts. Food courier riders overwhelmingly men Food couriering in Africa is overwhelmingly dominated by men. This is consistent with global trends where only one in ten workers on gig-based platforms are women. In some countries, app-based delivery platforms are an important source of work opportunities for migrants (ILO, 2021). In the three cities, only 2% of our sample were women. They were reluctant to talk to us. In Johannesburg, we could not help noticing the control some male riders had over the one woman rider we interviewed. She was never alone during the interview. The men drivers would listen in to what she was saying, and while interpreting in their language what she was saying, seemed to be telling her what she should say. This indicated that not only is this line of work maledominated, it also quite patriarchal. The fact that the woman worker was not comfortable to talk to us when the men riders were there indicated the dominance they exerted over her – and perhaps over other women riders at this particular work zone. Moreover, in Nairobi, the woman rider interviewed said she was attracted to the job because it was work in which“you can manage your own time”. She added, “Many ladies have high ratings and customers trust them.” However, safety is a major concern for women. She mentioned how robbers followed her one night but she managed to escape(Microsoft Teams: Masikane, 9 June 2021). Similarly, in Johannesburg women riders also spoke of safety as one of their major concerns in this line of work, a reason why many women choose not to do it. The challenges women riders face would be valuable research – not only around safety but also how men colleagues suppress their freedom of expression. It is worth noting though that the few women riders who joined the courier forum(see p. 19) have grown steadily in confidence. By the third forum, they were participating actively in the discussions and two women were chosen to sit on the eight-member interim forum committee. Food courier riders tend to be young: 59% of the total sample was under the age of 30. In Accra, as many as 82% of the riders were under the age of 30. Only 9% of the total sample was over 40 years of age. Photograph 4:“I have to carry this[pepper spray] when I do deliveries at night.”Riders’ Photo Competition, Yeoville, Johannesburg, South Africa(September 2021) by Cebile Mkhabela. 21

100% 80% 81.6% (n=40) 60% 40% Number of respondents(n)=149 46% 44% (n=23) (n=22) 50% (n=25) 34% (n=17) 100% 90% (n=45) 80% 60% 40% Number of respondents(n)=150 48% 46% (n=24) (n=23) 54% (n=27) 40% (n=20) 20% 18.4% (n=9) 0% 0% (n=0)(n=0) 0% Accra/Ghana Riders ages<30 30-39 10% (n=5) 0% (n=0) Johannesburg/South Africa 40-49 50-59 Graph 2: Riders’ ages by city and country Source: Webster and Masikane(2021) Education attained 14% (n=7) 2% (n=1) Nairobi/Kenya The riders tended to be well-educated. Of the total sample, 73% of respondents said they had matric 33 or above, and 27% held a tertiary education qualification. However, within the sample there was a sharp difference between Accra, where 38% had only a primary school certificate, and Nairobi, where 58% of the riders had a post-matric or higher qualification. Riders’ marital status Interestingly, with regard to marital status(Graph 3), 90% of the Accra respondents described themselves as single; this could be because they were young. Just over half(54%) of the Nairobi riders said they were married; they were middle-aged or older than the riders in Accra and Johannesburg. In Johannesburg, 48% of our sample described themselves as single. However, riders’ definition of‘single’ was based on their status of being ‘alone’ in Johannesburg. Of the riders who responded that they were single, 8% said that they were married back home. For example, one rider responded,“I am married at home, but single here.” The significance of riders across the cities being single was that they could work long hours without any family responsibilities; 33 Grade 12 in South Africa, and the equivalent of 0-Levels in Accra and Nairobi 34 For further investigation of this theme read: Social Reproduction Theory: Remapping Class, Recentering Oppression, by Tithi Bhattacharya, 2017. 20% 0% 10% (n=5) 0% 0% 0% (n=0)(n=0)(n=0) 0% 0% (n=0)(n=0) 6% (n=3) Accra/Ghana Riders marital status Johannesburg/South Africa married single widowed divorced 2% 2% 2% (n=1)(n=1)(n=1) Nairobi/Kenya coGraph 3: Riders’ marital status by city and country Source: Webster and Masikane(2021) being single reduced the costs of supporting a family. It also meant that they sell their labour at a cheaper price than local citizens who must provide for the cost of supporting themselves and their families. 34 However, some of the cross-border migrant couriers establish permanent relationships with local women; indeed, some even get married, as shown outside a home where a food courier wedding was in progress(see Photograph 5). Photograph 5:“Attending a wedding of a fellow friend and food courier rider.” Riders’ Photo Competition, Orange Grove, Johannesburg, South Africa(September 2021) by Justice Tafamba. Rent or own: riders’ accommodation As shown in Graph 4, an overwhelming majority(94%) of riders in all three cities rented their accommodation, because they were unable to buy a house or get a loan to buy a house. However, in Nairobi, riders seemed more settled, older, better educated and most likely 22

to be married. Fourteen percent(14%) owned their accommodation. In Accra, the riders were younger and not yet able to buy a house. One rider said,“I share my accommodation with four other riders”(Thywill, 2021, Accra). In Johannesburg, we found a similar situation, with riders sharing a communal house with four or five other riders. This further highlights the fact that the lack of stable income makes it difficult for riders to get a loan. Furthermore, the cross-border migrant status of some riders may mean they do not have the correct documentation to get a loan from the bank to buy a house. Number of respondents(n)=141 100% 97.6% (n=41) 98% (n=48) 86% (n=43) 80% 60% 40% 20% 14% (n=7) 2.4% (n=1) 2% (n=1) 0% Accra/Ghana Johannesburg/South Africa Rent or own accommodation own rent Nairobi/Kenya Graph 4: Riders’ accommodation: rent vs own Source: Webster and Masikane(2021) Riders’ length of employment as courier riders With regard to length of employment as a courier rider, the surprising finding – as Graph 5 reveals – is that employment was relatively stable: of the total sample, 82% of respondents had been in the job for over a year; while 44% had been in the job for more than three years. In the case of Johannesburg, 50% had been in the job for three years or more, while a surprisingly high 69% of the Nairobi respondents had been in the job for over three years. Indeed, in Nairobi as many as 12% had been in the job for more than five years. Our findings suggest that the courier riders are a relatively stable workforce, staying on in the job for relatively long periods of time. In 35 See Julie Zollmann,“Is Digital Driving Good Work in Nairobi?” Working Paper(Tufts University, March 2021) for further development of this argument. the case of Accra, where the riders were much younger and stayed in the job for shorter periods of time, the job could be a stepping-stone to another career. However, in Nairobi, our results show that food courier rider work may well be the equivalent of a career. 35 Number of respondents(n)=148 100% 80% 60% 67.3% (n=33) 57.15% (n=28) 40% 20% 0% 20.4% (n=10) 14% (n=28) 10.2% (n=5) 2% (n=1) 32% (n=16) 18% (n=9) 26.5% (n=13) 4.1% (n=2) 12.2% (n=6) Accra/Ghana Johannesburg/South Africa Nairobi/Kenya No. of years working<3 months 1-3 years 3-5 years 5-10 years Graph 5: Length of riders’ employment as a courier Source: Webster and Masikane(2021) In Johannesburg, the most likely explanation for staying on in the job, is that courier work was the best job available. As respondents remarked,“I just want to survive.” As migrants, this approach to work is more likely to enable riders to return to their home countries from time to time. This question requires further investigation. Which company do you ride for? Table 1 shows a breakdown of the tech giants for whom our sample of food couriers rode. , Table 1: Who do you ride for? 23

24

RESEARCH FINDINGS PART 2: THE NATURE AND WORKING CONDITIONS OF A BIKER’S JOB In this section of the report, we explore our findings around the nature of the job of a food courier biker and their working conditions. These key questions and themes emerge: 1. Riders’ perception of employment relationship: a worker or self-employed partner? 2. What attracted riders to working as food couriers? 3. Was being a food courier rider riders’ only occupation? 4. Riders who owned vs. rented their motorbikes 5. Whether the nature of the relationship between the riders, the courier company and the customer was neutral or biased 6. How important was the rider rating system to the riders? 7. The link between the rating system and riders’ commission 8. Hours of work and unpaid labour 9. Problems couriers faced at work 1. RIDERS’ PERCEPTION OF EMPLOYMENT RELATIONSHIP: A WORKER OR SELF-EMPLOYED PARTNER? As Graph 6 demonstrates, only 22% of the total sample of riders saw themselves as employees. This is particularly interesting to note because of how this line of work is presented as a self-employment or partnership opportunity. The majority of riders saw themselves as self-employed, as evidenced from our data: in Accra (76%) and Nairobi(79.6%). We discovered that only 18% of riders in Johannesburg described themselves as ‘self-employed’ and 56% of riders described themselves as‘partners’. When exploring this further, we interpreted their understanding of the term‘partners’ as being similar to self-employed. Johannesburg respondents describing themselves as‘self-employed’(18%) and ‘partners’(56%) – together 74% – was very close to those in Accra and Nairobi who preferred to see themselves as‘self-employed’. Number of respondents(n)=149 100% 80% 76% (n=38) 79.6% (n=39) 60% 10% (n=5) 40% 20% 24% (n=12) 20% (n=10) 18% (n=9) 20.4% (n=10) 0% 0% (n=0)(n=0) 0% 6% (n=3) 0% (n=0) 0% (n=0) Accra/Ghana Johannesburg/South Africa Nairobi/Kenya Worker or self-employed? employed self-employed partner contracted Graph 6: Riders’ perception of employment relationship: a worker or self-employed partner? Source: Webster and Masikane(2021) In Accra, some of the riders explained that the nature of their employment was“not under any control of an employer”(Abn, 2021). Another rider remarked,“I am not accountable to anyone”(Bright, 2021, Accra). Michael quipped,“I work for a third party, my boss” (Michael, 2021, Accra). Musala said he was happy that“Jumia gives you freedom to do other work!” (Musala, 2021, Accra). A Nairobi respondent remarked, “I am self-employed because I own my bike”(Makena, 2021, Nairobi). Riders’ responses go to the very heart of the legal dispute over whether they are independent contractors, that is, self-employed, or employees. 25

A dispute took place in 2018 when Uber B.V. 36 deactivated the Uber drivers’ app. The drivers claimed unfair dismissal and took Uber B.V. to the Commission for Conciliation, Mediation and Arbitration(CCMA) in South Africa. Uber South Africa(SA) rejected the CCMA’s jurisdiction on the basis that workers were independent contractors and not employees – see Uber South Africa Technology Services(Pty) Ltd v National Union of Public Service and Allied Workers(NUPSAW) and Others(2018) 37 and a law firm’s commentary 38 on this novel dispute. In their case, the drivers made reference to Section 200A and Section 213 of the Labour Relations Act(LRA), contending that they were employees. Workers argued this because they were required to perform their duties personally; customers created contracts with Uber and not the drivers as individuals, and drivers are largely controlled by Uber monitoring their performance by algorithm. Uber’s case was that drivers were not under any obligation to use the Uber app or drive Uber-registered vehicles; drivers could choose where to drive and which passengers to collect, and the drivers bore the risk of profit vs. loss. Uber SA also claimed that no contractual agreement existed between them and Uber drivers, as there is only a contractual agreement between drivers and Uber B.V. The CCMA agreed with the applicants in the case – the drivers. But the decision was overturned on a technicality when Uber SA appealed to the Labour Court. The Labour Court held that the CCMA failed to consider the fact that Uber SA and Uber B.V. are separate, independent entities and applicants should have also lodged a complaint against Uber B.V. Uber B.V. provides the legal contracts, technology, and deal with the collection and payment of monies received from the Uber drivers. However, Uber SA – being Uber B.V.’s local subsidiary – hires, controls and approves the Uber drivers. The Uber drivers predominantly, if not exclusively, engage with Uber SA daily and not Uber B.V. The Labour Court made it clear that the question of whether drivers were independent contractors or employees was left unanswered. It stated that the CCMA’s decision was solely overturned on a technicality: applicants should have also brought Uber B.V. to the CCMA when lodging their dispute (ibid., p. 33). On 19 February 2021, the UK’s Supreme Court ruled in favour of Uber drivers who argued that they were workers, not independent contractors Uber B.V. and Others vs. Aslam and Others. The court ruling stipulated that while drivers have fewer job rights than employees, they are provided with more benefits and protection than independent contractors. This ruling catalysed further attempts to legally challenge the employment status of Uber drivers and riders in South Africa. A few days after the UK ruling, a South African law firm, Mbuyisa Moleele Attorneys(MMA), indicated that they were preparing to launch a class action lawsuit against Uber B.V. MMA were supported by and collaborated with Leigh Day, the British law firm that represented some of the UK Uber riders and drivers. Although there has been no indication of the progress MMA and Leigh Day have made since announcing their intention to launch their initiative, it is likely their line of argument will be similar to Leigh Day’s in the UK’s Supreme Court case. Therefore, the central argument will revolve around the high levels of control Uber exercises over its riders. 2. WHAT ATTRACTED RIDERS TO WORKING AS FOOD COURIERS? Here we turn to the question of what attracted our respondents to becoming motorbike riders in the food courier business. Inevitably, responses included the attraction of‘freedom’. Graph 7 highlights key reasons that riders gave. 36 Uber B.V. is a Netherlands-based company. Consequently South African courts do not have any jurisdiction over it. 37 http://www.saflii.org/za/cases/ZALCCT/2018/1.html[16 July 2020]. 38 https://www.pagdens.co.za/legal_employment_status_of_uber_drivers/#:~:text=Uber%20SA%20objected%20to%20 the,of%20the%20Labour%20Relations%20 Act.&text=they%20are%20largely%20controlled%20by%20Uber[5 June 2020]. 26

Number of respondents(n)=145 40% 20% 38.8% (n=33) 22.4% (n=19) 22.4% (n=19) 16.5% (n=14) 40% (n=20) 34% (n=17) 30.1% 29% (n=28) (n=27) 22% (n=11) 21.5% (n=20) 19.4% (n=18) 4% (n=2) 0% Accra/Ghana Johannesburg/South Africa Nairobi/Kenya Attraction to work better pay working conditions freedom only work available Graph 7: What attracted riders to work as food courier riders? Source: Webster and Masikane(2021) The main reason that the total sample gave for choosing to work as courier riders was that it was the only work available(29%). In the case of Johannesburg, 40% of our respondents gave this reason. A respondent from Accra put it simply:“It is the only work I could find”(Clement, 2021, Accra). The next highest reason given was better pay – 27% of the total sample. This was highest in Accra(39%). Interestingly, greater‘freedom’ in the job was only the third highest reason given by the total sample(25%). In Accra, 17% mentioned freedom, 34% in Johannesburg, and only 29% mentioned this in Nairobi. As one rider remarked,“The app allows you to move around town instead of sitting at one place”(Thywill, 2021, Accra). In Johannesburg, a rider remarked,“There is no stress here. We work whenever we want with no one on our backs looking for us”(Takundwa, 2020, Johannesburg). However, the‘freedom’ couriers experience is somewhat ironic; riders are free to choose to work long hours in their own time, yet they are subject to the company’s algorithm which directs them where to go and how much to charge. If riders fail to satisfy, they can be disconnected and receive no income at all. 3. WAS BEING A FOOD COURIER RIDERS’ ONLY OCCUPATION? Related to workers’ choice of work, we asked respondents whether being a food courier rider was their only occupation. Their responses are captured in Graph 8. Number of respondents(n)=150 100% 80% 82% (n=41) 86% (n=43) 82% (n=41) 60% 40% 20% 18% (n=9) 14% (n=7) 0% Accra/Ghana Any other occupations? Johannesburg/South Africa yes no 18% (n=9) Nairobi/Kenya Graph 8: Was being a food courier rider workers’ only occupation? Source: Webster and Masikane(2021) Contrary to the idea of the gig economy as short-term work of a transient nature, 83% of the total sample of respondents said that being a food courier was their only occupation. Working exclusively as food courier riders was highest in Johannesburg(86%), followed by Accra and Nairobi(both 82%). However, of the total sample, 16% of riders said they do have side hustles, while 5% described themselves as students. In Johannesburg, 7% of riders have side hustles as mechanics or carpenters, or run a small business, such as a laundromat. In Nairobi, 6% of the riders were students. None of the riders in Johannesburg were students. 4. RIDERS WHO OWNED VS. RENTED THEIR MOTORBIKES Graph 9 shows that 58% of our total sample owned their bikes. However, there were regional variations: ownership was much higher in Nairobi(76%), whereas in Accra it was 57%, and 42% in Johannesburg. 27

80% 60% 40% 57.1% (n=28) 42.9% (n=21) 20% Number of respondents(n)=150 76% (n=38) 58% (n=29) 42% (n=21) 24% (n=12) 0% Accra/Ghana Johannesburg/South Africa Bikes owned or rented? owned rented Nairobi/Kenya Graph 9: Riders who owned vs. rented their motorbikes Source: Webster and Masikane(2021) Of the riders who owned their bikes, a total of 12% in all three cities owned more than one bike. Those riders who owned more than one bike, rented them to other riders to earn extra income. The Uber business model is premised on the assumption that riders own their bikes and that the company is simply offering drivers a technological platform to turn their idle asset into a productive resource(see Kute et al., 2020). The low proportion of riders who own their bikes is at odds with the intended benefits associated with Uber and other forms of gig-work platform work, particularly around the independence that bike ownership supposedly provides the riders. The CEO of Customer First Logistics in South Africa, Lawrence Dlomo, confirmed that most food courier riders cannot afford to buy a new bike. Instead Dlomo’s company, in partnership with Hyundai, rents brand new motorbikes to riders at R600(US$39) per week (Lawrence Dlomo, 30 September 2020). In return, he services the bikes and pays for insurance. Dlomo said the fact that most of the riders are undocumented means that, in the event of an accident, he cannot claim from the insurance company. He alleged that riders have fake licences and are unable to open bank accounts in their own names. We also asked those riders not riding their own bike, who owned the bike they rented. In 20% of cases it was a tech giant company, such as Mr Delivery, or in 19% of the cases, an individual or an independent owner. An example of an independent owner is Mobile Macs. In a few cases, the bike was owned by the family(3%) or by a friend(9%). It seemed that often someone owned several bikes and rented them out to couriers as a way of making extra money. A substantial number of our respondents (44%) indicated that they had paid for the bike from their savings; a minority of respondents(12%) reported having to take out a loan. Table 2 indicates estimates from riders for the cost of buying a bike. Table 2: Estimated costs of a buying a bike Bike servicing, as we can see in Table 3, is another cost in the courier business. US$ Table 3: Estimated costs of servicing a bike per month Table 3 captures data about bike servicing, from our sample of respondents: • A total of 28% of riders across the three cities said they serviced their bike weekly; 37% serviced their bike every two weeks, and 35% serviced their bike monthly. • Of the total riders who serviced their bike weekly, 43% were in Accra, 35% in Johannesburg, and 6% in Nairobi. • Of riders who serviced their bike every two weeks, 31% were in Accra, 39% in Johannesburg, and 40% in Nairobi. • Of the riders who serviced their bike monthly, by city: Accra(25%), Johannesburg(27%), and Nairobi(54%). 28

29

The servicing of bikes indicates that riders take care of their bike on the one hand; on the other, frequent servicing says something about the condition and quality of the bikes they use. The condition of their bikes could further speak to the number of accidents that the riders have reported. Number of respondents(n)=148 60% 54% (n=27) 40% 20% 42.9% (n=21) 30.6% (n=15) 24.5% (n=12) 38.8% (n=19) 34.7% (n=17) 26.5% (n=13) 40% (n=20) 2% (n=1) 0% 0% (n=0) 6% (n=3) 0% (n=0) Accra/Ghana Johannesburg/South Africa Nairobi/Kenya Servicing of motorbikes every week every 2 weeks monthly never Graph 10: How frequently do riders get their bikes serviced? Source: Webster and Masikane(2021) Motorbike insurance Motorbike insurance is another vital part of being a courier rider, with 69% of participants reporting that their bikes were insured. An overwhelming 96% of the riders in Nairobi had their bikes insured; 4% of participants did not know whether their bikes were insured because they did not own them. In Accra, 92% of the riders’ bikes were insured. By contrast in Johannesburg, only 20% of riders said they had bike insurance, 48% of the riders’ bikes were not insured, and a further 32% of riders who did not own their bike, did not know if the bike was insured. 39 5. IS THE NATURE OF THE RELATIONSHIP BETWEEN THE RIDERS, THE COURIER COMPANY AND THE CUSTOMER NEUTRAL OR BIASED? We asked our rider sample about their perception of the relationship between themselves, the courier company and the customer. Table 4 shows that almost half of the respondents saw it as a neutral relationship. Total Table 4: Is the relationship between the riders, the courier company and the customer neutral or biased? We found interesting regional variations, as can be seen in Graph 11, in the relationship perception: the majority of riders in Nairobi(60%) and Accra(47%) saw the relationship between the company, themselves and the customer as neutral, whereas half(50%) of the riders in Johannesburg perceived the relationship as being unbalanced, in favour of the company and the customer. A Columbine Square, Johannesburg respondent said, “They side with the client because they get money from them”(Benjamin, Clearwater Mall). Another rider said,“The customer is always right... they say”(Alex, Campus Square). Surprisingly, a far larger percentage of riders(27%) in Accra saw the company as being more on the side of the workers than those in Johannesburg (2%) and Nairobi(2%). Number of respondents(n)=148 60% 46.65% (n=21) 50% (n=25) 60% (n=30) 40% 38% (n=19) 26.65% (n=12) 28% (n=14) 20% 17.8% (n=8) 8.9% (n=4) 10% (n=5) 10% (n=5) 2% (n=1) 2% (n=1) 0% Accra/Ghana Johannesburg/South Africa Nairobi/Kenya Perception of bias neutral biased towards workers biased towards customers NA Graph 11: The nature of the relationship between the riders, the courier company and the client Source: Webster and Masikane(2021) 39 As part of her research strategy of participant observation, this report co-author, Masikane, registered as a food courier rider with Uber Eats in Johannesburg, South Africa. At no stage in the registration process was the question of insurance raised as part of the requirements. 30

6. IS THE RIDER RATING SYSTEM IMPORTANT TO YOU? The vast majority(83%) of our respondents said that the service rating system was important to them. Riders get feedback via the app on the quality of their service by way of a stars rating system. The lowest rating score is 1 star, the highest, 5 stars. The rating scores are added to give the rider and the customer an idea of how many stars the rider has overall. For example, a rider would have about 4,000 stars and this is reflected on the riders’ details. Riders said when they know how their service is received by their customers, it motivates them to improve their service.“It pushes drivers to comply because they want higher ratings,” said Josias(Rosebank Mall, Johannesburg, 2020). A small percentage of respondents did not think the rating system was important to them: Accra(6%), Johannesburg (16%), and Nairobi(18%). By introducing an element of competition through rating and providing visible evidence of success, the courier riders are drawn into ‘playing the game’. Sociologists call it‘gamification’, and it is at the centre of the Uber business model. Sarah Mason, 40 who took a job as an Uber driver to study the work model, described it in this way:“Simply defined, gamification is the use of game elements – pointscoring, levels, competition with others, measurable evidence of accomplishment, ratings and rules of play – in non-game contexts. Games deliver an instantaneous, visceral experience of success and reward, and they are increasingly used in the workplace to promote emotional engagement with the work process, to increase workers’ psychological investment in completing otherwise uninspiring tasks, and to influence, or‘nudge’ workers’ behaviour”(Mason, 2018:4). Why is rating important? Most of the participants believed that customer ratings were important because rating affected their chances of getting potential clients.“When you get poor ratings from clients, Uber Eats deactivates 40 Sarah Mason was a doctoral student at the University of California at the time. 41 Uber, 2016, Driver Deactivation Policy: Sub-Saharan Africa Only, September. See: https://www.uber.com/en-ZA/blog/ driver-deactivation-policy/[Accessed 9 August 2020]. your account,” said Joseph(2020, Campus Square, Johannesburg).“When it[your rating] is as low as 85 they[Uber Eats] send you a message warning you about your ratings. If there are no improvements, they deactivate your account.” Riders’ commission is a crucial reason why rating is important; it provides riders with a steady income, and it motivates them. The commission is the amount that Uber Eats would charge per trip. When asked why the driver rating was important, riders explained that a high rating was important in ensuring their continued access to the Uber network and, by extension, a good rating helped protect their income stream generated on the app. Each city on Uber’s global network has its own minimum rating level that drivers must maintain for them to stay active on the platform. 41 This provides riders with an instrument to control income and further solidify the perceived‘flexibility’ they enjoy. Here are some riders’ comments as to why the rating system is important to them: “Bolt Foods uses ratings to give more orders to a person with good ratings”(Edward, 2021, Accra). “Bolt Foods uses the rating system to reward hardworking riders occasionally”(Bright, 2021, Accra). “Bolt Foods gives rewards, like bonuses, occasionally when you have high ratings”(Iddrissa, 2021, Accra). “It works as a system because riders with good ratings always get orders. There is always work for the riders with good ratings”(Mensali, 2021, Accra, referring to Bolt Foods). “The App automatically rates the riders at the end of the month. It gives categories – gold, bronze, copper and stars. This is done monthly, and the riders share the ratings with Solar Taxi – the Company”(Samuel, 2021, Accra). “I have learnt a lot about how to handle customers since I started this work. My human relation has improved, and I have come to know a lot of places too”(Christopher, 2021, Accra). “They[Uber Eats] value the client’s comments so much. And the clients will rate you according to their moods” (Lucas, 2021, Nairobi). 31

In Johannesburg, riders generally thought that the rating system worked well. One rider said that ultimately,“It is a business. Someone must complain, and you need to know how your service is being received, you have to make a plan”(Benjamin, 2020, Johannesburg). 7. THE LINK BETWEEN THE RATING SYSTEM AND RIDERS’ COMMISSION Answers to our question about whether riders knew how much commission their app company was charging revealed sharp differences between the cities, as depicted in Graph 12. In Johannesburg, 86% of riders said they did not know how much commission the company was charging them. In Accra, 61% of riders did not know, whereas in Nairobi only 4% did not know. Number of respondents(n)=149 80% 86% (n=43) 76% (n=38) 60% 61.2% (n=30) 40% 30.6% (n=15) 20% 0% 4.1% (n=2) 4.1% (n=2) 0% (n=0) 8% 6.3% (n=4) (n=3) 0% 0% (n=0)(n=0) Accra/Ghana Johannesburg/South Africa 12% (n=6) 4% (n=2) 6% (n=3) 2% (n=1) Nairobi/Kenya Knowledge about amount of commission charged 20-25% 30% don’t know fixed amount no commission Graph 12: Riders’ knowledge about how much commission they were being charged Source: Webster and Masikane(2021) Of those riders who did know the amount of commission charged: In Nairobi, 76% of the riders said the commission was between 20% and 25%; 12% of riders said the commission charged was under 30%; and 6% of riders said there was a fixed and standard commission amount. 42 42 A number of the riders did not know this answer and consequently the Nairobi data did not add up to 100%. In Accra, 4% of the riders said the commission charged was between 20% and 25%; 31% of riders said the commission was 30%; and 2% said there was a standard and fixed amount. In Johannesburg, 6% of riders said the commission was between 20% and 25%; and 8% of riders said it was 30%. 8. HOURS OF WORK AND UNPAID LABOUR In our total sample of respondents, 45% work 10 or more hours a day, and often a six-day week. In Johannesburg, 69% work 10 hours a day – a 60-hour week. Because remuneration is task-based, the more tasks – the more deliveries – riders undertake, the more money they make. The riders refer to these long hours as an attempt at[reaching your] target. In other words, if the driver reaches a specific amount of money in a week, they can at least cover their daily expenses, such as rent, petrol and groceries. In the words of a Johannesburg courier,“It’s up to you. If you want to meet your target you work more. If you want to make extra you must work harder” (Chris, 2020, Campus Square, Johannesburg). How long do you wait between calls? On average, in the total sample, 63% of riders waited for between 30 minutes and one hour between calls. But in Johannesburg, 71% waited for only 15 minutes. In addition to waiting for an order, some couriers complained of waiting for clients to come and fetch their orders when they deliver. But this seemed a minority view. A Ghanaian courier remarked,“Eighty percent of my clients are responsible people and the restaurants are also very good. Before you get to the restaurant, the order is ready and there is no waiting for long”(Felix, 2021, Ghana). Length of working time Forty-six percent(46%) of the total sample said the length of the average working day was from 9am to 9pm. As a Ghanaian courier rider remarked,“I do not have off days. I work all days of the week. I only pause 32

when there is an occasion”(Richard, 2021, Accra). In Johannesburg, a higher proportion(74%) of the couriers said they work from 9am to 9pm. Eighty percent(80%) worked for over 56 hours a week, but there was strong regional variation. In Johannesburg, riders seem to work for longer hours compared to Nairobi. Graph 13 illustrates the impact of long hours on riders’ family time. Number of respondents(n)=149 60% 65.3% (n=32) 62% (n=31) 50% 50% (n=25)(n=25) 40% 30.6% (n=15) 20% 22% (n=11) 16% (n=8) 4.1% (n=2) 0% Accra/Ghana Johannesburg/South Africa Impact on family time not affected affected take time off 0% (n=0) Nairobi/Kenya Graph 13: Impact of riders’ working hours on family time Source: Webster and Masikane(2021) 9. PROBLEMS COURIERS FACED AT WORK In this section, we begin with riders’ income. Do you earn normally enough to cover your monthly expenses? Thirty-seven percent(37%) of the total sample of riders said they earned enough to cover their monthly expenses. The highest proportion of riders who said they earned enough was in Accra(48%), followed by Nairobi (39%) and then Johannesburg(25%). Although 32% of riders said they could not cover their monthly expenses, a significant 31% said they did not know. The high number of‘don’t knows’ is likely a result of the fact that riders’ earnings are unpredictable because income depends on how many orders they receive every month. As one rider in Accra remarked,“I live within whatever income I am able to make in the month[enough that] I am able to pay rent and utilities”(Yussif, 2021, Accra). But clearly for many riders their work was a hand-tomouth existence as these two comments attest:“Some deliveries are calculated based on distance, so the income is not adequate”(Ali, 2021, Accra); and“My salary is not enough. I have siblings in school and parents to take care of. My dependants are more than my salary” (Emmanuel, 2021, Accra). Joseph, a food courier in Johannesburg, spoke about the difficulties of estimating your income because he does not have a fixed income: “I know how the system works but sometimes we have challenges logging in. Sometimes one driver gets more trips than the other. We do not know how it works. Data only lasts for a week. If you go on Facebook and WhatsApp[to communicate with family and friends] your data will finish quicker. The Uber app takes up too much data. One GB of data is enough if you do not chat or use Facebook. One day, you wake up and make R200[US$12.77]. Some days we make more and others less. There’s no fixed amount that we get paid per day”(Joseph, 2020, Campus Square). If you do not cover expenses, where do you get the extra money? Family, friends and neighbours(24%) seemed to be the most common source of filling riders’ expenses gap, but the largest category was other informal economic activities. Not surprisingly, few drew on their savings: Nairobi and Johannesburg were both 4%. A significantly higher number of riders in Accra(23%) drew on their savings. Graph 14 captures additional problems riders faced. The most common problems riders faced in all three cities were: • Police harassment(22%) • Unfair clients(20%) • Stress(19%) • Accidents(15%) • Illness(12%) • Crime(11%) 33

Number of respondents(n)=388 30% 20% 10% 22% (n=29) 28.8% (n=38) 20.4% (n=27) 16.7% (n=22) 9.1% (n=12) 23.3% (n=24) 20.4% (n=21) 15.5% (n=16) 14.6% (n=15) 9.7% (n=10) 11.7% (n=12) 3% (n=4) 4.8% (n=5) 0% (n=0) 0% Accra/Ghana Johannesburg/South Africa Problems at work stress illness crime police harassment unfair clients 18.9% (n=29) 17% (n=26) 22.9% (n=35) 18.9% (n=29) 9.2% (n=14) 13.1% (n=20) 0% (n=0) Nairobi/Kenya accidents uncoded 7 Graph 14: Workplace problems facing riders Source: Webster and Masikane(2021) In ranking riders’ responses by the most significant, by city, crime topped the list in Johannesburg(57%). In Accra, 45% of riders cited police harassment, whilst in Nairobi police harassment was at 41%. The surprisingly low figure of 12% of Johannesburg riders citing police harassment may be related to corrupt Metro Police accepting bribes. In an informal conversation with couriers, bribery was accepted as a norm, with a modest amount of R20 per incident(US$1.40). Next are some of the riders’ responses on the nature of workplace problems. In all three cities, accidents were a major problem facing food courier riders. Photograph 6:“Praying for God’s protection from accidents.” Riders’ Photo Competition, Hillbrow, Johannesburg, South Africa (September 2021), by Justice Tafamba. “During COVID-19 hard lockdown period, I was affected heavily. I could not cater for my needs. Meanwhile I had developed a lot of back pains due to working long hours”(Sakim, 2021, Nairobi). “Riders are not provided with PPEs[Personal Protective Equipment], but they expect riders to buy their own PPE. Even phones being used by riders are charged against their salaries”(Alfred, 2021, Accra). “I just received a message from a client that they have COVID, I should not enter their house. This shows that many clients are sick, yet Uber Eats has not provided us with any PPE”(A conversation between a rider and Fikile Masikane on WhatsApp, 2021). “The company is not treating us well. They think about their money and their motorbike. When you are sick, involved in an accident, or something happens and you are not able to work, the company is not happy with us”(Felix, Solar Taxi, 2021, Accra). Other riders express more mixed responses: “Jumia pays on time. I have had three accidents since I started working in delivery work. I have learnt how to relate well with customers”(Sherif, 2021, Accra). “So far it has been helpful to work in the delivery service work, I have been a bit financially stable” (Bright, 2021, Accra). “Delivery work is a decent occupation, and I will tell other young people with no jobs to join”(Richard, 2021, Accra). “Generally, delivery service has been good for me. I can make a decent living out of it. A key problem worth mentioning is the treatment from our dispatchers at Jumia – refusal to schedule us on time, or sometimes not at all. They have little patience for us when problems arise”(Michael, 2021, Nairobi). “Bolt needs to find a way to license riders to avoid police doubts, by giving the riders a[company] logo for the bikes”(Okio, 2021, Accra). 34

35

RESEARCH FINDINGS PART 3: WHAT POWER DO FOOD COURIERS HAVE TO CHALLENGE THE UNEQUAL WORK RELATIONSHIP? In this section we explore what potential the food courier riders have to get the tech giant companies to respond to their demands, including opportunities to organise. Graph 15 shows data about riders’ daily place of work. 100% 93.6% (n=44) Number of respondents(n)=147 98% (n=49) 80% 60% 74% (n=37) 40% 20% 0% 6.4% (n=3) 0% (n=0) Accra/Ghana 16% (n=8) 10% (n=5) Johannesburg/South Africa 2% (n=1) 0% (n=0) Nairobi/Kenya Place of work shopping centre/mall smartphone both Graph 15: Where is your daily place of work? Source: Webster and Masikane(2021) As indicated in the graph, most riders(88%) saw their smartphone as their daily place of work. As one respondent remarked:“Without my phone, I have no work”(Steven, 2021, Clearwater Mall, Johannesburg). In Johannesburg, 26% of riders saw both the shopping mall and the smartphone as their place of work. This reflects the high number of malls and supermarkets in this city. Relationship with co-workers We asked riders whether they had a relationship with their co-workers. Although geographically isolated, the majority(88%) said they developed relationships with their co-workers. Country by country, 19% in Accra said they had not, followed with a similar answer by 10% of riders in Johannesburg, and 8% in Nairobi. Trade unions and social networks Regarding whether there were trade unions or unionlike organisations for riders, of the overall sample, only 12% said yes, 46% said no, while 46% of riders said they didn’t know. One driver said,“I would like to be unionised as we take 100% risk. We are exposed to diseases and are often robbed. Nyaope boys take our phones every day; some people rob us”(Ronald, 2020, Southgate Mall, Johannesburg). While most of the drivers confirmed there were no trade unions among food delivery riders, one of the drivers said that the WhatsApp groups operate like a trade union. A rider gave the example of the Brothers of Melville(BOM), an informal society of food courier riders at Campus Square in Melville, Johannesburg. The chairperson of BOM is quite clear on their demands. They want to be permanent drivers for several reasons. He believes there is negligence in Uber Eats, rider fatalities, work insecurity, and very little support.“They do not give you anything, besides the paper bag when you go to collect an order”(Joseph Apio, 2020, Campus Square, in-depth interview). 36

For these reasons, this rider started two main WhatsApp groups: BOM and United Ugandans(UU). Because of the individualistic nature of the job, it is difficult to connect with drivers from elsewhere, so they all have separate groups; the Congolese and the Zimbabwean riders have their own groups as well. Unless riders work at Campus Square together, or are from the same country, it is difficult to connect. Photograph 7: The rules and regulations of the Brothers of Melville (BOM), Campus Square, Johannesburg, South Africa(March 2020) by Fikile Masikane. The groups, BOM and UU, have rules and regulations (see Photograph 7). If a rider does not agree with them), they are free to leave. These groups function as support groups. They contribute money and share valuable information. For example, they each contribute R100(US$6.40) when something happens, such as when a member is involved in an accident, dies, or is the victim of robbery. They also collect R20 (US$1.20) every Sunday for informal savings and split the money between them at the end of the year. There are about 20 people in each group. One of the members of BOM said,“These groups help us. When you are alone you cannot go far, especially in South Africa, you always need support from people. It is better to be part of these groups”(Alex, 2020, Campus Square). In Accra, 9% of riders said they were members of a group; in Johannesburg, 25%; and in Nairobi, 21%. Graph 16 shows the breakdown of different reasons why riders joined WhatsApp groups. 80% 87% (n=40) Number of respondents(n)=136 60% 40% 44.1% (n=15) 32.4% (n=11) 44.6% (n=25) 25% (n=14) 28.6% (n=16) 20% 13% (n=6) 0% (n=0) 0% 0% 0% (n=0)(n=0) Accra/Ghana 2.9% (n=1) 14.7% (n=5) 5.9% (n=2) Johannesburg/South Africa 1.8% 0% (n=1) (n=0) Nairobi/Kenya Why riders jpoin whatsapp groups informal savings info about work info about working conditions other all Graph 16: Reasons riders joined a food couriers’ WhatsApp group Source: Webster and Masikane(2021) The purpose of riders’ WhatsApp groups was to share information about work(59%) and working conditions (20%) – this represents a total of 79% of riders’ sharing information on their work. Informal savings was the reason mentioned by 11% of the sample. Of the Nairobi cohort, 25% mentioned informal savings. This is likely the result of the widespread formalisation of saving groups, such as the Savings and Credit Cooperative Organization (SACCO). 43 Photograph 8:“Our leader addressing us in a meeting.” Riders’ Photo Competition, Parktown, Johannesburg, South Africa (September 2021) by Justice Tafamba. 43 They(riders/boda-boda drivers) register the union with the Ministry of Cooperatives, which in turn authorises SACCO to receive deposits and provide loans to its members. SACCO is run by members who are selected by the other members. The emergence of these cooperatives was a response to the long bureaucratic processes associated with Kenyan banks. 37

EXTRACT FROM FIKILE MASIKANE’S FIELD NOTES JOURNAL F ood couriers are not only organised at work. They also come together after working hours, especially on public holidays, to participate in their own soccer tournaments. It was intriguing to see riders relaxing in a different setting. At the invitation of Mr Delivery rider, Brian Chirwa, Edward Webster and Fikile Masikane visited Pieter Roos Park, Parktown in Johannesburg on Freedom Day, 27 April 2021, to watch their soccer tournament. possibility of someone raping, hijacking or robbing her while she was doing evening deliveries in certain areas. But she said she continued with being a food courier rider because she needed the work. When I asked her about her relationship with her co-workers, she smiled and said,“These are my‘brothers’ and some are my ‘children’. They take good care of me. They protect me all the time. I can rely on them. We have a good relationship.” She further explained how she ran the riders’ WhatsApp group. When someone had an issue, for example, a road crash, she would take care of cooking and buying groceries for the injured driver. In this way, Londi plays a nurturing role for the riders. Photograph 9: Food courier riders’ bikes parked at the Pieter Roos Park, Parktown, Johannesburg, South Africa(April 2021) by Fikile Masikane. We arrived at 1pm and were met by Brian who was warmly welcoming and keen to talk. Brian is a member of the Food Couriers Forum. We wanted to record an updated interview with him, since he was part of the study at its inception in March 2020. We were joined later by Lovemore Ndlovhu and Thando Dlodlo from Zimbabwe, who both worked for Uber Eats. Later, Londiwe Kumalo, one of the few women riders from Uber Eats joined us. Of all the people who attended this soccer match, Londiwe and I were the only women present. She and I had an informal conversation about her experiences as a woman courier in a male-dominated space. She expressed how challenging it was, especially when it came to issues of safety at work. She said she feared the Photograph 10: Londi Kumalo, one of the few women food courier riders interacting with the players at the soccer match at Pieter Roos Park, Parktown, Johannesburg, South Africa(April 2021) by Fikile Masikane. The purpose of the soccer tournament was for the riders to take a break from their busy work schedules. Gatherings such as these offer an important opportunity to connect with one another in a relaxed setting. They had taken the afternoon off because, Brian said, most public holidays are not as busy because customers are generally at home and prefer to cook their own food. The riders sat around in groups, talking and laughing. They all seemed very friendly towards each other. The courier riders – mainly from Checkers Sixty60, Uber Eats and Mr Delivery – had parked their bikes together. The initial communication stated that 30 riders would be in attendance, but to Brian’s surprise – considering that most drivers would rather work than take some time off – 50 drivers showed up for the soccer tournament. It took them a while to get started but eventually they divided into six teams of seven each, and began the afternoon’s 38

competition. The winning team was awarded a trophy from contributions gathered from all the riders. Photograph 11: The riders being divided into groups for the final game. Pieter Roos Park, Parktown, Johannesburg, South Africa(April 2021) by Fikile Masikane. As the tournament continued, I realised that a few riders had brought their children along to watch, taking the opportunity to have family time, since they do not see each other as often due to work pressure. The riders were quite happy to be playing soccer while interacting with each other about work-related issues. During the match Zakhele, one of the riders who scored a goal for Uber Eats, was eagerly embraced and slapped on the back by many of the riders. Londi told me that Zakhele had been badly injured in a bike accident a few months back. The riders’ excitement was mainly because they were so happy that he was still alive and by then fit enough to play soccer. This speaks to safety at work, one of the main issues riders in our research complained about. From our observations we drew these general insights: • The courier riders are already well organised and even have a photographer. They are not organised into a trade union, although they don’t seem hostile to the idea. They come together face-to-face. They use their phones to communicate, but a few riders clearly take the lead. They see themselves as having common interests as an occupational group. As a rider said:“We are all facing the same problems, so we come together to help each other”(Lovemore, 2021, Johannesburg). • All the couriers were from Zimbabwe, except for one man from Uganda and one woman from South Africa. • None of the couriers from outside South Africa have work permits and seem to think permits are impossible to obtain. When the police stop them, riders give them R20, and the police let them continue on their way. But their lack of work permits seems to be one of the key problems couriers face. • Riders value very highly the freedom they get from not having a direct boss; they can work whenever they want to, and, for example, take time off to take their children to school. But the couriers also want some security and compensation for accidents. One courier at the tournament was still limping from a fractured ankle caused by a road crash. • Interestingly, the riders reacted to the idea of being regulated. They seemed to feel that regulation would jeopardise their already precarious situation. Although their work is dangerous, they earn a lot more than security guards. The couriers seemed quite proud of what they do. • The couriers are well educated; and some want to continue with their studies. • They are determined to make a better living in South Africa. Photograph 12: McDonald riders vs Uber Eats riders at a soccer tournament. Riders’ Photo Competition, Pieter Roos Park, Parktown, Johannesburg, South Africa(September 2021) by Justice Tafamba. 39

60% 45.5% (n=5) Number of respondents(n)=150 59.2% (n=29) 47.9% (n=23) 40% 27.3%27.3% (n=3)(n=3) 29.2% (n=14) 20% 0% 0% (n=0) (n=0) 0% ( Accra/Ghana Reasons for strike action 18.4% (n=9) 18.4% (n=9) 4.1% (n=2) 0% (n=0) 12.5% (n=6) 8.3% (n=4) 2.1% (n=1) Johannesburg/South Africa Nairobi/Kenya payment safety commission police harassment/other NA Graph 17: Rider strikes: What were the demands? Source: Webster and Masikane(2021) The most common demand riders had was around payment, followed by police harassment. Safety was not a high demand; it did not appear in our sample in Accra, and in Nairobi and Johannesburg it appeared in only 2% and 4% of responses, respectively. A major food courier riders’ strike with Uber Eats in Cape Town in 2018 was over the company cutting rates. The strike was unsuccessful because some riders from other work zones did not know about it. Apps can also isolate drivers and riders, as they do not reveal how many drivers and riders there are, or who they are. However, in terms of organising, riders are often scared to reveal their identities, as they are fearful of being algorithmically punished or kicked off the app. Since most food courier riders in Johannesburg are crossborder migrants, they are worried that by participating in acts of collective resistance they would increase their chance of being discovered by the company, such as Uber Eats or Mr Delivery, and get deactivated. A meeting spot at Campus Square allows the Johannesburg riders to share their common grievances regarding the nature and the conditions of their work. As courier rider Brian states,“The poor working conditions are all we ever talk about when we meet”(17 March 2020 interview). As discussed earlier this led to online solidarity and the formation of WhatsApp groups, such as the Uganda Bike Drivers Association. Although this association is not formally registered as a trade union, it functions as a type of trade union through acts of collective solidarity. Alex, a fellow Ugandan courier, recalled an occasion when a car drove into one of the bikers from Uganda and the biker was killed(17 March 2020 interview). The group immediately went to assist. They helped with the funeral arrangements and the return of the rider’s body to Uganda. There is no joining fee for the group; money is only collected on a case-bycase basis. This WhatsApp group has 60 members, all from Uganda. Brothers of Melville(BOM) in Johannesburg, reported on earlier, organised a strike in February 2020. They demanded an increase in the amount of money they earned for each trip. They said the rate was too low compared to the kilometres they drove for each trip. They used their WhatsApp group to alert all drivers about the strike. If one driver was found working on the roads during this time, they took away his phone.“We take away your office,” Alex a member of BOM said. Due to the strike, Uber Eats was losing money per trip, as riders refused [to take over] the trip.“We withdrew our labour and we refused to take any trips. We collectively refused to take trips. Uber Eats did not victimise anyone”(17 March 2020 interview). Many rider participants in our research were reluctant to respond to questions concerning strikes, but of those who did, 44% said there had been rider strikes in response to their grievances, with a high incidence in Nairobi(68%) and Johannesburg(58%). Of those who said there had been strikes, 63% said there had been as many as three strikes. First city-wide strike of platform workers in South Africa 44 On 18 December 2020, in response to a fare decrease, roughly 2,000 courier riders across Johannesburg collectively logged off the Uber Eats platform, forcing 44 This article was researched and written by Jamie Rosengarten, a research assistant on this project and an undergraduate student at the University of the Witwatersrand.[3 February 2021] 40

the company to halt operations across the city. Scores of couriers in the big cities of Tshwane, Durban and Port Elizabeth also collectively logged off. Under pressure during the COVID-19 pandemic lockdown on restaurants, Uber Eats had lowered the fare they were required to pay from a 30%(plus 5% to the rider) commission per order, to a flat rate of approximately R9 (US$0.57) per meal. This change meant a loss of potential income for the rider. To cover the losses associated with these pricing changes, on Monday 14 December 2020, without consulting any riders, Uber Eats indicated their intention to lower the commission rate paid to couriers. While courier rider commission rates vary in different geographical areas, couriers claimed they were paid about R4(US$0.26) per km. This was only the latest of many reductions in their fares, with courier riders being paid almost half the rate they were paid three to four years ago. While localised food courier rider protests are relatively common occurrences, and generally ignored by management, a city-wide demonstration was widely viewed both by management and trade unions as something that isolated, independently-contracted riders did not have the organisational capacity to achieve. The courier riders released a memorandum of seven demands. They wanted the following: • Increased delivery fees • Safer routes • The arbitrary suspension of courier accounts stopped • Improved safety and provision of better in-trip support • Labour brokers(company set up to supply labour to clients) banned • Delivery bags and Uber Eats equipment supplied • The ability to choose not to accept cash trips. The courier riders’ primary demand was increased fares. They organised strikes using WhatsApp groups. This enabled the strike organisers to overcome the company platform attempts to isolate them. Strikers set up communication networks and channels from which they could spread messages and communicate across multiple WhatsApp groups. Strikers were unable to achieve their demands because Uber Eats was unwilling to engage with them. They rejected the strikers’ demands. This meant both the initial and the one-day follow-up strike on 22 January 2021 failed to achieve the riders’ immediate intention of increasing income and improving levels of security. The strikers demonstrated their ability to overcome the hyper-individualised identities that platforms impose on them. They were able to organise coordinated provincial and nation-wide resistance to Uber Eats. Previous strikes and collective logouts were very regionally-specific. Uber Eats easily stopped them by refusing to collectively bargain with couriers. However, never had courier strikes received as much media attention as these strikes received. With extensive coverage in the media for the first time, a food courier riders’ strike in South Africa was televised on a TV news platform. Although the strikers’ demands were not met, they extensively developed the power resources from which to capitalise on in the future. 41

BUILDING WORKER ORGANISATION FROM BELOW We turn in our last section to the possibilities of new forms of food courier workers’ organisations emerging from below. The emergence of union-like associations Our research findings point towards the emergence of union-like organisations existing side-by-side with traditional trade unions to defend workers’ needs and interests in the digital economy. In other words, those who speak of the‘end of labour’,“are speaking of the end of a particular kind of worker organisation, in particular the traditional industrial union”(Visser, 2012). 45 Instead, what we see emerging in the digital economy are hybrid forms of organisation, including different types of associations that blur the distinction between traditional unionism and informal workers’ associations or cooperatives, as illustrated in Diagram 4. 46 “Unions are reinventing themselves,” argues the Indian Federation of App-Based Transport Workers(IFAT). They are doing this by“utilizing the newest technology to connect with the workers and collaborate with independent researchers and knowledge-based research organizations”(IFAT, 2020:5). As Gadgil and Samson(2017) argue in their study of a trade union of waste pickers in India, “increasing numbers of informal worker organisations are developing hybrid forms to address a range of needs far wider than those of formally employed wage workers”(Gadgil& Samson, 2017:162). They describe the politics of developing hybrid organisations as complex and dependent on the recognition“that the union form can be simultaneously crucial and insufficient for meeting the needs of their members”(ibid., p. 160). Trade Unions Self-organised networks HYBRID FORMS OF WORKER ORGANISATIONS Associations Worker centres/ NGOs Cooperatives Diagram 4: Hybrid forms of worker organisations Source: Webster, Ludwig, Masikane and Spooner(2021) Traditional unions remain crucial in providing support and access to institutional power for the emerging organisations of precarious workers(Webster et al., 2021). It is misleading to characterise this approach as trade union fetishism as Atzeri(2020) does, because it assumes that the organisational form of trade unions is static. 45 “The particular confluence of sector and occupation that held together one sector=one union idea of the‘industrial union’... was 30 or 40 years ago common in European countries but is today the exception”(Visser, 2012:130); “The expansion of the public sector and the unionisation of teachers, nurses, and many other service providers in the welfare state have led to even the mainstream and left-leaning federations becoming, for the most part, white collar organisations”(Visser, 2012:135). 46 Visser(2012) identifies five different types of unions in the industrialised world: the traditional industrial union that organises blue collar workers in a particular sector covering all skill grades including technicians, supervisors, and white collar staff; a narrow industrial union that only organises non-manual workers or white collar staff in a particular sector; the classical blue collar craft union; the occupational white collar union or staff association that organises particular occupations and professions irrespective of sector; general unions that cut across occupations and sectors(Visser, 2012:136). 42

This is demonstrated most clearly in the response by the Transport and Allied Workers Union of Kenya(TAWU-K), and the shift in their organising strategy to recruit and organise platform workers. See the highlighted panel, for an example. Organising digital platform workers in Kenya 47 The Transport and Allied Workers’ Union of Kenya (TAWU-K) is a registered union affiliated to the Central Organisation of Trade Unions(COTU) and the International Transport Federation(ITF). After the Uber business model disrupted the traditional taxi labour market in 2018, under the direction of its digital organiser Bill Mutoro, TAWU-K began to organise drivers in the platform economy.“It became clear,” Mutoro argued,“that the drivers needed representation, voice and collective bargaining.” After an initial survey, TAWU-K discovered that the drivers had already organised themselves into associations and a variety of societies. In these early organising endeavours, these informal initiatives came across numerous challenges as they relied on traditional organising strategies. Among their challenges were: • Riders’ reluctance to pay union dues made it costly to recruit workers • A weak democratic culture – or a lack thereof – in these emerging structures • The union’s constitution did not accommodate representation of associations and informal workers. As a result of these challenges, TAWU-K embarked on a new organising and recruitment strategy designed specifically for platform workers. This was done in four key steps: Step 1: Restructure the union to accommodate platform workers. Step 2: Develop new organisational tools and devise new ways of collecting fees. Step 3: Explore ways of challenging the law through litigation. The union realised that the biggest stumbling block was the disguised nature of the employment relationship. Drawing on Convention No. 198(2006) of the International Labour Organisation(ILO) on employment relationships, TAWU-K decided to focus on the nature of work done, as opposed to the employment relationship and their mis-classification of platform workers as independent contractors. Step 4: Focus on the sector rather than the company level. Because many riders have multiple platformbased apps and are often working for different companies, it was decided to develop collective bargaining at a centralised sector level for all platform workers. Because Uber and Bolt deny that they are transport companies and claim to be service companies, it became clear that changes in the Kenyan legislation would be necessary. Furthermore, these companies argue that they are not subject to national law, because their headquarters are abroad. Based on this new organising strategy, the union managed to recruit over 2,000 drivers, and expanded its activities from Nairobi to Mombasa, Nakwu, Kisumu, Edoret and Mt Kenya. An important part of its current strategy is to build membership in Mombasa through promoting the idea of a social protection platform. One of the most innovative aspects of their organising strategy is the goal of developing their own software product. The product will enable them to send automatic message confirmation notifications every time they receive payment from union members, as well as dispatch general communications to union members. In developing their strategy TAWU-K is guided by Recommendation 204(2013) of the ILO on transitioning from the informal to the formal economy. Source: TAWU-K.“Organizing digital labour platforms in the transport sector: collective action, litigation and legislation.” Prepared by Bill Mutoro 2021. 47 The information in this panel is drawn from a report by Bill Mutoro, digital organiser for the Transport and Allied Workers Union of Kenya(TAWU-K) and from an interview with the authors on 8 August 2021. 43

The Power Resources Approach(PRA) The case of TAWU-K suggests that through internal restructuring and new organising strategies, unions can meet the needs of informal workers and the new workers of the digital economy. To meet these needs, unions will have to transform and rediscover their power and the capabilities necessary to realise this power. The framework for such a task can be drawn from the Power Resources Approach(PRA). The intellectual foundations of the PRA were laid by Erik Olin Wright(2000) and Beverly Silver(2003) over two decades ago. The two key concepts – structural power, the power stemming from labour’s position in the economic system; and associational power, the power arising from collective political or trade union workers’ associations – provided the basis for the approach. In the decade that followed, labour scholars went on to identify two other sources of workers’ power, namely institutional and societal power(Chun, 2009; Schmaltz, Ludwig& Webster, 2018; Schmaltz, Ludwig& Webster, 2019). The concept of symbolic power – or societal power – was added into the PRA by researchers in the United States. They argued that workers with limited structural power were able to compensate for the lack of associational power,“by drawing upon the contested arena of culture and public debates about values” (Chun, 2009:7). Researchers from Germany discussed the role of institutions for labour power, arguing that organised labour can draw upon institutional power resources, such as institutionalised labour rights and institutionalised dialogue procedures; these are sources of power that labour can rely on even when structural and associational power is weakened (Dörre, Holst& Nachtwey, 2009; Urban, 2013; Schmalz & Dörre, 2013). In addition to these debates on the nature of labour power, scholars from Canada argued that specific capabilities are needed to mobilise the individual power resources(Lévesque& Murray, 2010; 2013). The Power Resources Approach is illustrated in Diagram 5. STRUCTURAL/ ECONOMIC POWER Marketplace bargaining and workplace bargaining power ASSOCIATIONAL POWER Stability/vitality of unionisation INSTITUTIONAL POWER Securing in influence institutional SOCIETAL POWER Coalitional power and discursive power Diagram 5: Trade union power resources Source: Schmalz, Ludwig and Webster, 2018:120 If effective regulation of platform capitalism is to emerge, then collective representation and voice for workers in the digital economy will be necessary. Vandaele(2018) points us in the right direction when he says that,“stopping machines in the twentieth century corresponds to collective logouts in the twenty-first century”(Vandaele, 2018:15). He then goes on to stress the constraints facing digital workers going on strike, such as management’s ability to ‘disconnect’ and the difficulties of coordinating collective action. But he suggests three ways of enhancing digital workers’ bargaining power through applying the Power Resources Approach. This can be illustrated by drawing a distinction between actual and potential power. Firstly, digital workers can create worker-driven messaging apps and chat groups where they can share information, develop a shared identity, and announce local direct action. Direct interaction at their work zones contributes“to their self-organisation and associational power in the making”(ibid., p. 16). As we illustrate in the slide diagrams on pages 46-48, riders have the potential for workplace bargaining power through being technologically-linked through the app. 44

45

STRUCTURAL POWER Workplace bargaining power ACTUAL POTENTIAL Riders have weak workplace bargaining power. Riders, because they are technologically-connected, have the potential for bargaining power by implementing collective logouts. Firstly, although their classification as employees weakens their associational power, the potential for associational power is drastically increased when workers meet face-to-face at work zones and begin to form a collective identity. ASSOCIATIONAL POWER ACTUAL POTENTIAL Riders’ major weakness is that they are not classified as employees, but have an intermediate level of associational power. Riders have social media group chats, but are fragmented between nationalities and languages. Potential for riders’ associational power could be dramatically expanded if riders: • Meet at work zones and begin to form a collective identity. • Develop campaigns around social protection. • Had centralised avenues of communication between group chats. • Could form union-like structures to announce local direct action, attract media attention, and forge a shared identity, trust and solidarity. Secondly, platform workers can form alliances with trade unions or other organisations, who can assist them by taking test cases to courts, providing financial assistance to strikers, or gaining support from the providers(the restaurants) or the clients(the customers). 46

SOCIETAL POWER ACTUAL Riders have relatively weak societal power: alliances with unions are limited. POTENTIAL • Shape public discourse by debunking management’s narrative about entrepreneurship. • Win public sympathy. • Form alliances with unions and labour-oriented NGOs. Thirdly, digital workers can influence the public discourse by debunking management’s narrative about entrepreneurship, delegitimise the platforms’ employment practices, and build discursive or social power. This can translate into rulemaking and institutional power through setting minimum standards on wages and social protection. INSTITUTIONAL POWER ACTUAL Riders have weak institutional power. POTENTIAL Riders could use a combination of structural and associational power to translate it into institutional, policy and legislative change. However, the major limitation of couriers is that they have weak marketplace bargaining power. Entry requirements for the job are low and large reserves of unemployed labour make riders easily replacable. Riders in Johannesburg are especially vulnerable as they are largely cross-border migrants and can be deported. A union strategy that requires management to limit the number of riders operating at a time would undoubtedly improve riders’ marketplace bargaining power. 47

STRUCTURAL POWER Marketplace bargaining power ACTUAL Riders have very weak marketplace bargaining power because: • In Johannesburg, riders are cross-border migrants. • The job has low entry requirements. • There are large reserves of labour and riders are easily replaceable. • Companies disconnect‘disorderly’ riders. POTENTIAL A union strategy limiting the number of riders operating at a time would undoubtedly improve riders’ marketplace bargaining power. Essential capabilities to be developed Food couriers’ actual power is weak. It will require the development of their capabilities for their bargaining potential to be realised(Levesque& Murray, 2013). Four essential capabilities can be identified. The ability to: • Learn from the past(learning capabilities) • Resolve conflict and build consensus (intermediation) • Develop new strategies and define an autonomous agenda(framing) • Adapt organisational traditions to changes in policy needs. We have argued in this report for the development and exercise of power by platform workers. Simply put, power is the ability of A to get B to do something B otherwise would not have done. But as Wright argues, the outcome of A exercising their power may not necessarily be against B’s interest(Wright, 2000). 48 This definition of power considers the potential mutual benefit of both persons involved in the conflict. It is, as Wright argues, the difference between a negative compromise and a positive compromise. A negative compromise is one where each party can inflict considerable costs on the other, but neither is capable of decisively defeating the opponent. A positive compromise is one where the contending forces find a way to actively cooperate in ways that open some space for non-zero-sum gains. 48 Of course Wright is not denying that a conflict of interest exists between capital and labour; he is simply arguing that there will be occasions when there are common interests. 48

LOOKING TO THE FUTURE The disruption caused by the digital age and platform businesses exists in a regulatory vacuum. This presents both a challenge and an opportunity to the governments of African countries, the companies, and the people who work in these platforms. There has been an important initiative to rate platform companies by an organisation called Fairwork. Its goal is to show that better and fairer jobs are possible in the platform economy. The Fairwork project is based at the Oxford Internet Institute and the WZB Berlin Social Science Centre. Through its global network of researchers, they evaluate the working conditions of digital platforms and rank them based on their five principles of fair work: fair pay; fair conditions; fair contracts; fair management; and fair representation. The principles were developed through a collaborative process that reflects the insights of their international network and workers’ voices around the world (Fairwork, 2021). However, it is not clear to what extent Fairwork has become proactive in designing regulatory frameworks for the platform economy. 49 In the absence of an adequate regulatory framework for platform work, two broad pathways can be identified: the first involves a deepening of the domination of foreign-owned tech giants with no national or global agreement on how to operate. This will create some informal jobs, but workers will be stuck in low-wage drudgery with none of the protections or benefits of formal employment. With profits and taxes retained abroad, this could be described as a form of re-colonisation of the global South(Couldry& Mejias, 2019). An alternative pathway could be a‘digital social compact’, created with the active participation of platform workers and their organisations. This would involve coherent global and national policies, including legislation to protect such workers. This optimistic path opens the possibility of the extension of labour and social protections to informalised workers. In the latter scenario, the new technology of platform work creates opportunities – what Chacaltana and Leung(2019) call‘entry points’ – that facilitate access to social protection, simplify registration, and support compliance with laws. It can, they suggest, create a trend to e-formality. This alternative developmental path can best be achieved through social dialogue among the relevant stakeholders, most particularly the digital labour platforms, the platform workers and their representatives, and relevant government departments. The first step in this process is for the many union-like structures emerging in the gig economy to come together and to create a national coordinating structure. 49 Katta et al.(2020) argue that the COVID-19 pandemic has destabilised Uber’s ‘conjuctural’ existence and forced the company to become more embedded in the locations where it operates, bringing about a – perhaps temporary – turn towards the decommodification of its drivers’ labour. 49

50

ACKNOWLEDGEMENTS We thank our partners FES-SA and TUCC for their support throughout the research phase. In particular, we thank Bastian Schulz, the director at FES-Kenya who started this research with us in 2020, followed by his successor Kathrin Meißner, the director of FES-SA in 2021, Shane Choshane the project manager, Masego Mofokeng, and the TUCC team. We thank the FES-Kenya office, particularly Brenda Chelagat and Marion Mukasi for facilitating the process of getting a fieldwork researcher. We thank Agnes Tsheri, our fieldwork researcher in Nairobi and Bill Mutoro at TAWU-Kenya for his keen assistance in supporting our work. In Ghana, we thank Ebow Mensah from the FESGhana office for his contribution and Karim Saagbul, who was our fieldwork researcher in Accra. We could not have done the comparison without their assistance, especially with the COVID-19 pandemic restrictions. We would also like to thank SCIS for their support and their contribution to the Riders’ Photo Competition. We would like to thank Pat Horn and Dinga Sikwebu for their contributions at the rider forums in Johannesburg. We thank Michael O’Donovan, a data analyst who assisted us with the comparative statistics of the three cities. We also thank Jamie Rosengarten for the riders strike contribution in December 2020. In particular, we would like to thank Karen Hurt for her very professional editorial work, Nadia Sinovich for graphic design and Marcus Toerien for copy-editing and proofreading. Most importantly, we thank and acknowledge the riders in Johannesburg, Accra and Nairobi who generously allowed us to spend time with them over the past two years. As you continue to fight for your rights, we encourage unity and remember that an injury to one is an injury to all! REFERENCES Anner M, Pons-Vignon N& Rani U. 2019. For a Future of Work with Dignity: A Critique of the World Bank Development Report, The Changing Nature of Work. Global Labour Journal, 10(11), pp. 2–18. Anwar MA& Graham M. 2020. Between a rock and a hard place: Freedom, flexibility, precarity and vulnerability in the gig economy in Africa. Sage, 25(2), pp. 237–258. Online at https://doi. org/10.1177/1024529420914473. Atzeri M. 2020. Worker Organisation in Precarious times: Abandoning trade union fetishism, Rediscovering Class. Global Labour Journal, 11(3), pp. 311–314. Balaram B, Warden J& Wallace-Stephens F. 2017. Good Gigs: A fairer future for the UK’s gig economy. London: RSA(Action and Research Centre). Agency staff. 2020. Amazon boss quits in solidarity with fired workers. Business Day, 5 May 2020, pp. 11. Cant, C. 2020. Riding for Deliveroo: Resistance in the new Economy. Cambridge: Polity Press. Castel-Branco R, Mapukata S& Webster E. 2020. Work from home reserved for the privileged few in South Africa. Business Day, 17 September 2020. Online at https://www.businesslive.co.za/bd/ opinion/2020-0917-work-from-home-reserved-for-the-privileged-fewin-sa/. Chacaltana J, Leung V& Lee M. 2018. New technologies and the transition to formality the trend towards e-formality. Semantics Scholar. Chau D. 2019. Deliveroo sued by former employee who alleges exploitation and underpayment. ABC News, 28 August 2019. Online at https://www.twu. com.au/companies/deliveroo/deliveroo- sued-by-formerworker-who-alleges-exploitation-and-underpayment/ 28 August 2019. 51

Chun JJ. 2009. Organizing at the Margins. The Symbolic Politics of Labor in South Korea and the United States. Ithaca, NY: ILR Press. Conger K& Scheiber, N. 2019. California Bill Makes App-Based Companies Treat Workers as Employees. New York Times, 11 September 2019. Online at https:// www.nytimes.com/2019/09/11/technology/californiagig-economy-bill.html. Cossyleon J& Spitz G. 2021. Collaborating, then stepping back: Doing public sociology through participatory research. In Hossfield L, Brooke Kelly E& Hossfield C(eds), The Routledge International Handbook of Public Sociology. London and New York: Routledge, pp. 69–78. Couldry, N& Mejias U. 2019. The Costs of Connection: How data is Colonizing Human Life and Appropriating it for Capitalism. Stanford University Press. Elliot J. 2013. Suffering Agency: Imagining Neoliberal Personhood in North America and Britain. Social Text, 115 Vol 31 No. 2 Summer. Duke University Press. Fairwork South Africa, Ratings 2021 : Labour standards in the gig economy. Fleming. 2017. Human Capital Hoax: Work, Debt and Insecurity in the Era of Uberization. Organization Studies, 38(5), pp. 691–709. Frey CB& Osborne M. 2013. The future of employment: How susceptible are jobs to computerisation? Oxford: Oxford Martin School Publications. Friedman G. 2014. Workers without employers: Shadow corporations and the rise of the gig economy. Review of Keynesian Economics, 2(2), pp. 171–188. Gandini A. 2019. Labour Process theory and the gig economy. Human Relations, 72(6), pp. 1039–1056. Global Labour Institute(GLI). 2020. Kampala Bus Rapid Transit: Understanding Kampala’s Paratransit Market Structure. Manchester: GLI. Govindarajan V& Srivastava A. 2018. Reexamining dual-class stock. Business Horizons, 6(3), pp. 461– 466. Huws U. 2016. Logged labour: a new paradigm of work organisation? Work Organisation, Labour& Globalisation, 10(1), pp. 7–26. Huws U. 2014. Labor in the Global Digital Economy: The Cybertariat Comes of Age. New York: Monthly Review Press. Gadgil M& Samson M. 2017. Hybrid Organisations, Complex Politics. When Unions form Cooperatives. In Webster E, Britwum AO& Bhowmik S(eds), Crossing the Divide. Precarious Work and the Future of Labour. Scottsville: UKZN Press, pp. 143–164. Gandini A. 2019. Labour process theory and the gig economy. Human Relations, 7(6), pp. 1039–1056. Indian Federation of App-Based Transport Workers(IFAT). 2020. A study about the response to the pandemic by the App-based companies: New Delhi: IFAT in collaboration with the International Federation of Transport Workers(ITF). International Labour Organisation(ILO). 2021. World Employment and Social Outlook: The role of digital labour platforms in transforming the world of work. Geneva: ILO. Internet World Statistics(2021): Internet Penetration in Africa 2020- Q1. Online at https:// www. internetworldstats.com/stats1.htm[Accessed 2 April 2020]. Johnson C, Dunn M.& van Vuuren P. 2020. Digital platforms’ role in African digitisation and gig work on the back of COVID-19, April 21 2020. Online at https:// cenfri.org/articles/digital- platforms-role-in-africandigitisation-and-gig-work-on-the-back-of-covid-19/. Katta S, Badger A, Graham M, Howson K, Ustel, S& Bertolini B. 2020.(Dis)embeddedness and(de) commodification: COVID-19, Uber, and the unravelling logics of the gig economy. Dialogues in Human Geography, 10(2), pp. 203–207. Kute S, Molife M& Tshezi M. 2020. Workers or Partners? The Political Economy of Uber in Dar es Salaam, Nairobi, and Johannesburg. Friedrich Ebert Stiftung(FES). Levesque C& Murray G. 2013. Understanding Union Power: Resources and Capabilities for Renewing Union Capacity. European Review of Labour and Research, 16(3), pp. 333–350. Lewin K. 1946. Action research and minority problems. Journal of Social Issues, 2(4), pp. 34–46. 52

Lozano AA. 2018. Reframing the Public Sociology Debate: Towards Collaborative and Decolonial Praxis. Current Sociology, 66(1), pp. 92–109. Mason S. 2018. High score, low pay: why the gig economy loves gamification. The Guardian. Online at https://www.theguardian.com/business/2018/nov/20/ high-score-low-pay-gamification- lyft-uber-drivers-ridehailing-gig-economy. Mastercard Foundation. 2019. Digital Commerce and Youth Employment in Africa. Nocke V, Peitz, M& Stahl K. 2007. Platform Ownership. Journal of the European Economic Association, 5(6), pp. 1130–1160. Online at https://doi. org/10.1162/JEEA.2007.5.6.1130[Accessed 12 May 2020]. Rahman KS& Thelen K. 2019. The Rise of the Platform Business Model and the Transformation of Twenty-First Century Capitalism. Politics& Society, 47(2), pp. 177–204. Schmalz S, Ludwig C& Webster E. 2018. The Power Resources Approach: Developments and Challenges. Global Labour Journal, 9(2), pp. 113–134. Srnicek N. 2016. Platform Capitalism. Cambridge and Malden: Polity Press. Silver B. 2014. Theorizing the Working Class in 21st Century Global Capitalism. In Atzeni M. Workers and Labour in a Globalised Capitalism: Contemporary Themes and Theoretical Issues. London: Palgrave Macmillan, pp. 46–69. South Africa in the Digital Age(SADA). 2019. Pathways to Digital Work. Johannesburg: Genesis Analytics. Tassinari A& Maccarrone V. 2020. Riders on the Storm: Workplace Solidarity among Gig Economy Couriers in Italy and the UK. Work, Employment and Society, 34(1), pp. 35–54. Taylor M. 2017. Good work: Taylor Review of Modern Work Practices: July 2017. London: UK Department for Business, Energy& Industrial Strategy. Vandaele K. 2018. Will trade unions survive in the platform economy? Emerging patterns of platform workers’ collective voice and representation in Europe. Working Paper. Brussels: European Trade Union Institute(ETUI). Visser J.(2012). The Rise and Fall of Industrial Unionism. Transfer: European Review of Labour and Research, 18(2), pp. 129–141. Webster E& Bischoff C. 2011. New actors in employment relations in the periphery: Closing the representation gap amongst micro and small enterprises. Relations Industrielles/Industrial Relations, 66(1), pp. 11–33. Webster E. 2020. The Uberisation of work: the challenge of regulating platform economy: a commentary. International Review of Applied Economic, 34, pp. 512–521. Webster E, Ludwig C, Masikane F& Spooner D. 2021. Beyond traditional trade unionism: innovative worker responses in three African cities. Globalizations, 18(8), pp. 1363–1376. World Bank. 2018b. World Development Report 2019: The Changing Nature of Work. Washington DC: World Bank. Online at https://www.worldbank.org/en/ publication/wdr2019[Accessed 3 April 2020]. World Bank. 2019. Achieving Broadband Access for All in Africa Comes With a$100 Billion Price Tag. Washington DC: World Bank. Online at https://www. worldbank.org/en/news/press-release/2019/10/17/ achieving-broadband-access-for-all-in-africa-comeswith-a-100-billion-price- tag. Wright EO. 2000. Working-class Power, Capitalist-class Interests and Class Compromise. American Journal of Sociology, 105(4), pp. 957–1002. 53

“I just want to survive.” Publication Impressum Published by Friedrich-Ebert-Stiftung, Trade Union Competence Centre for Sub-Saharan Africa and the Southern Centre for Inequality Studies. Friedrich-Ebert-Stiftung, Trade Union Competence Centre for Sub-Saharan Africa 34 Bompas Road, Dunkeld West, 2196 Johannesburg, South Africa Tel:+27-10 446 8612 Email: tucc@fes.de © Friedrich-Ebert-Stiftung 2022 Southern Centre for Inequality Studies, University of the Witwatersrand North Lodge, Parktown Management Campus, 2 St David’s Place, Parktown, 2050 Johannesburg, South Africa Tel:+27-11 717 8312 ISBN: 978-0-620-98698-4(print) ISBN: 978-0-620-98699-1(e-book) The commercial use of all media published by the Friedrich-Ebert-Stiftung(FES) is not permitted without written consent. The publication may be quoted, and extracts used, provided that the source is fully acknowledged. The views expressed in the publication are not necessarily those of the Friedrich-Ebert-Stiftung. 54

55

Trade Union Competence Centre for Sub-Saharan Africa