TRADE UNIONS IN TRANSFORMATION 4.0  LABOUR AND SOCIAL JUSTICE GLOBAL LABOUR UNREST ON PLATFORMS The case of food delivery workers Vera Trappmann, Ioulia Bessa, Simon Joyce, Denis Neumann, Mark Stuart, Charles Umney September 2020 Over 500 incidents of labour unrest have been recorded globally within the last 3 years. Grass-roots organisations of riders are typically organising protests, with an increasing transnational coordination. Issues of pay, employment status and health protection have been the sources of discontent of workers, suggesting potential for closer collaboration between riders’ groups and trade unions.

FRIEDRICH-EBERT-STIFTUNG – Global labour unrest on platforms LABOUR AND SOCIAL JUSTICE GLOBAL LABOUR UNREST ON PLATFORMS The case of food delivery workers »Trade Unions in Transformation 4.0« examines unions’ strategic actions to mobilize power resources in a»new world of work« in which capital uses digital technology to re-organize the labour process. The Global Trade Union Programme of the FES aims to understand how the power balance between capital and labour is impacted and how workers are responding to the threats of the digital rollback towards greater exploitation and precariousness of workers. Pursuing a dialogue and action-oriented approach, the project ultimately intends to contribute to trade unions’ strategic reflections, experimentation and purposeful transformation.

Content Content 1 INTRODUCTION 2 2 THE RETURN OF ASSOCIATIONAL POWER 2 3 METHODS AND DATA 3 4 THE EXTENT OF LABOUR UNREST 5 5 POWER RESOURCES: ASSOCIATIONAL POWER AS THE MAIN SOURCE OF POWER 7 6 WHAT TYPE OF ACTION AND FOR WHAT GRIEVANCE? 9 7 CONCLUSION 10 References............................................................................................... 12 Annex....................................................................................................... 13 1

FRIEDRICH-EBERT-STIFTUNG – Global labour unrest on platforms 1  INTRODUCTION Labour unrest by platform workers is an important and growing phenomenon in the world of work. Yet we only have knowledge about individual cases of labour unrest and a more comprehensive picture is still lacking. Case studies point to various instances where platform workers have successfully defended their interests, mobilised or built coalitions, helping to understand suitable conditions for mobilisation(e.g. Cant, 2019; Tassinari and Maccarrone 2017), but»small n« case studies cannot tell us much about wider global trends in labour unrest in the platform economy. Little is known about the extent of labour unrest by platform workers, the forms such labour unrest can take or the key factors that animate such action. This study offers the first attempt to map such platform worker labour unrest on a global scale through the development of a unique index of labour unrest. Our analysis is of interest to both labour activists and scholars of employment relations and aims to offer some initial insights to a number of pressing questions. First, what kind of issues or grievances prompt platform workers to engage in bouts of labour unrest? Second, what types of actions do platform workers employ? Third, what actors are involved in different forms of platform labour unrest – platform workers themselves, established trade unions or emerging, grass-root organisations? Our initial, comparative analysis of platform labour unrest found that across different labour platforms offering services like cleaning, repair, care, childminding or delivery, labour unrest was most prevalent in the food delivery sector (Joyce et al. 2020). Building on this, we focus this report on the food delivery sector 1 , and examine patterns of platform labour unrest on a global scale, drawing from a database of over 500 instances of labour unrest in the sector. Our results show that labour unrest among food delivery workers has been growing in recent years and has spread across a large number of countries around the world. Conceptually, we locate our analysis with reference to the power resources approach(Silver, 2003), contributing to the»key task« of »identification of emerging responses from below to both the creative and destructive sides of capitalist development (ibd. 20)«. We argue that in order to understand patterns of labour unrest amongst platform workers, we need to revisit the concept of associational power. Specifically, we need to better understand how associational power is emerging and how it relates to other forms of power. We will show that action is widespread globally, and intensifying. We will show that labour unrest is driven by workers themselves, but increasingly receives support from trade unions. Issues driving unrest are not necessarily linked to novel characteristics of platform work, but more frequently to»classic« conflicts of distributional quality like pay and working conditions. In order to defend their interests, platform workers cannot rely on extensive power resources, but 1 We use food delivery here, but in the period studied, the companies included in our analysis mainly focused on cooked meal delivery, not on grocery deliveries. in the first instance mainly strive to establish associational power. The remainder of the report is organised in six sections. In the next section we outline key contributions to debates on power resources and consider how the concept of associational power can be applied to recent struggles by platform workers. Following this, we present the methodology behind our global index of platform labour unrest. We then present our quantitative findings linking the unrest to associational, institutional and societal power resources that platform workers might use or develop. In conclusion, we reflect on the key themes that emerge from our data and what this means for emerging forms of associational power amongst platform workers in the food delivery sector and more widely. 2  THE RETURN OF ASSOCIATIONAL POWER The power resources approach, as derived from Wright (2000) and Silver(2003), distinguishes between structural and associational power. Structural power refers to forms of power gained from the leverage workers have due to their situation within wider economic processes. It can be divided into‘workplace bargaining power’, denoting the extent to which workers are enmeshed in the process of production, and‘marketplace bargaining power’, which refers to the dynamics of supply and demand in a given labour market (for instance, workers have marketplace power if they possess scarce skills that are in high demand). Associational power is that power which comes from workers’ collective organisation, most notably through the organisation of trade unions. Silver(2003) contrasts the forms of union organisation and labour unrest characteristic of auto workers, who had high levels of structural power, with those of textile workers, who lacked structural power and consequently were far more reliant on associational power. Broadly speaking, Silver’s account, like the earlier work of Wright (2000), reflects the problem of structure and agency. That is, for Silver and Wright, structural power is about where workers are in the labour markets and productive processes of capitalism, whereas associational power is about what workers do to organise and represent themselves, and to further their interests. The distinction between structural and associational power has obvious practical implications for those interested in building(and rebuilding) effective union organisation. At a time when unions almost everywhere have suffered dramatic declines in size and influence, the attractions of a framework for understanding what capacities unions might mobilise to rebuild are obvious. As a result, the power resources approach has developed rapidly at the intersection of academia and activism. Two main directions of research can be identified. First, researchers and activists have looked for new sources of structural power upon which union revival might be built- hardly surprising, given the newly exposed weakness of unions’ associational power. This approach has 2

METHODS AND DATA led, for example, to a highlighting of transport and logistics as potential»choke points«(Alimohamed-Wilson and Ness 2018) in the»new terrain«(Moody 2017) of global production networks. Second, researchers have extended Wright and Silver’s initial framework to identify additional sources of workers’ power. Here, a non-exhaustive list includes:»symbolic power« (Chun 2005; Von Holdt and Webster 2008);»logistical power«(Webster 2015); combining structural and associational into»social power«(Bank Muñoz 2017); and, most pertinently for platform workers,»institutional power and societal power«(Brinkmann et al. 2008; Schmalz and Dörre 2013). Institutional power is usually seen as the result of struggles and negotiation processes based on structural power and associational power and manifests itself in legal institutional frameworks or collective bargaining systems or workplace representation(Dörre and Schmalz 2014). Societal power refers to the latitudes for action arising from viable cooperation contexts with other social groups and organisations, and society’s supports for trade union demands. It draws on»coalitional power« and»discursive power«; being able to create broader networks with other social groups and being able to successfully intervene in public debates(ibd.). As Vandaele(2018: 10) notes, given that platform workers’ institutional power is‘almost non-existent’,‘they need to rely on other resources for raising their voice and attaining bargaining power’. Their potential to form coalitions with other actors, or to exercise discursive power may therefore offer a more effective means of action, but this also suggests a need to(re-)focus attention on the processes of building associational power. There has been a recent revival of interest in associational power. Specifically, its importance has been emphasised by researchers investigating the struggles of precarious or informal workers in the global South – or in highly insecure employment in the global North – and often in the absence of trade union membership or organisation(Murray 2017; Doellgast et al. 2018; Eaton et al. 2017; Rizzo and Atzeni 2020). Where workers engage in militant, self-organised collective action, but where unions are not present, a broader understanding of associational power, which extends beyond union representation, seems appropriate(cf. Atzeni and Ness 2018). Moreover, for workers without formal employment rights, who consequently lack resources of institutional power, the extension of institutional coverage is often an aim of collective action rather than a pre-condition or an outcome(Doellgast et al. 2018). Again, this points to the need for a broad concept of associational power( what workers do) that can encompass emerging collectivities and grasp the nature and dynamics of labour unrest among highly precarious workers. We apply a broad concept of associational power here for understanding the significant and sustained wave of labour unrest and worker activism in global platform work. Platform workers’ unrest is one area where an implicit focus on associational power has been resurgent. A huge number of detailed case studies of platform workers’ organising activities(Anwar and Graham 2019; Brizziarelli 2018; Cant 2019; Chesta et al. 2019; Ford and Honan 2019; Jesnes et al. 2021; Kirk 2019; Minter 2017; Tassinari and Maccarrone 2017; 2019) have sought to understand the factors(Levesque and Murray 2007) that contribute to increasing associational power in the platform economy. An initial categorisation of the associational power of platform workers has been proffered by Vandaele(2018), which, following Offe and Wiesenthal(1980), assumes that grass-root unions and self-organising workers mainly pursue a logic of membership trying to mobilise and organise workers(thus increasing associational power), while more mainstream unions pursue a logic of influence trying to apply institutional power. This suggest that different actors pursue highly tailored kinds of organising activity depending on circumstances: they may utilise different kinds of strategies, depending on different kinds of partnerships, they may respond to different types of concern or grievance, or have differing attitudes towards institutional and legal functions. Vandaele(2018) usefully elaborates his categorisation of associational power with reference to key European examples of the successful collective organisation and representation of platform workers. Yet, we still lack insight into emerging patterns of associational power amongst platform workers across a wider geographical sweep. Further, key issues beyond the specific form of associational power remain relatively unexplored. First, just how extensive is platform labour unrest in a geographical sense, and is it possible to detect any comparative patterns of action? Second, what sorts of issues and grievances prompt platform labour unrest and how is this manifested in particular actions? Finally, what constellations of actors are responsible for the organisation of platform labour unrest, and how is this related to particular kinds of labour unrest. These issues are explored in the remainder of the report, drawing on an analysis of a global dataset of labour unrest in the food delivery sector. 3  METHODS AND DATA The study involved the collation of a unique dataset of platform labour unrest around the world, the so-called Leeds Index, which currently covers labour unrest in different platform sectors. In this study, the focus was on incidents of labour unrest experienced by companies in the food delivery sector. With the notion of labour unrest, we not only captured strike activity, but all sorts of activities that workers and collective actors used to advance workers’ interests and shift the power balance between labour and capital in their favour. These ranged from public demonstrations, campaigns, strikes and logoffs to legal court cases. There are six global food delivery brands – that is, brands operating on more than two continents or world regions(at the moment of data collection). These are Uber Eats, Just Eat, Deliveroo, Foodora, Zomato, and Glovo. Some of these brands are subsidiaries of bigger companies – for example, Foodora is part of Delivery Hero and Just Eat has merged with Takeaway. In addition to these six global players, we also included the two most important companies by geo3

FRIEDRICH-EBERT-STIFTUNG – Global labour unrest on platforms Table 1 Overview of selected food delivery platform companies Company HQ 1 Uber Eats 2 Glovo 3 Zomato 4 Just Eat Takeaway.com 5 Deliveroo 6 Foodora 7 Doordash 8 Jumia Food 9 Mr.D Food 10 Rappi 11 Pedidosya 12 Yandex.Eda 13 Delivery Club 14 Talabat/Carriage 15 Carriage 16 Yemeksepeti 17 Ele.Me 18 Meituan 19 Foodpanda Source: own elaboration. San Francisco, USA Barcelona, Spain Gurgaon, India London, UK London, UK Berlin, Germany San Francisco, USA Lagos, Nigeria Cape Town, South Africa Bogota, Colombia Montevideo, Uruguay Moscow, Russia Moscow, Russia Al Kuwayt, Kuwait Doha, Qatar Istanbul, Turkey Shanghai, China Beijing, China Berlin, Germany Significance Global Global Global Global Global Global USA, Canada, Australia Africa Africa South America South America Russia& Central Asia Russia& Central Asia Middle East Middle East Turkey China China Southeast Asia Number of countries the company is/was active in 3 47 29 24 16 13 9 2 9 1 9 10 6 7 10 6 1 1 1 12 graphic region(based on the United Nations geoscheme), with selection based on: the number of downloads of food delivery apps per country between February and April 2020 on Apple and Google Play Store; the number of countries the company was operating in per region; and revenues per company. 2 Our final list included Jumia Foods and Mr D. Food for Africa; Pedidosya and Rappi Doordash for the Americas; and Delivery Club, Yandex.Eda, Foodpanda, Talabat, Carriage, Meituan, Ele.Me. Talabat, Foodpanda and Carriage are subsidiaries of Delivery Hero. For Europe, no additional regional players were included, as global players dominate the landscape there. In sum total, these 19 food delivery companies have operations in 95 countries around the world. 3 Having identified 19 companies, we then searched electronic news archives for reports on events that mentioned their names. Two sources were utilised: China Labour Bulletin, mainly for China; and the GDELT project, which monitors worldwide news reports, with real-time translation of online news articles in over 100 languages, and a news search interface. The sources for GDELT are the world’s largest news agencies combined with the Google News algorithm. Although GDELT is known for its own machine-based event-coding system, we conducted our own searches for articles on each of the companies using relevant keywords capturing labour unrest, including:»protest«,»strike«,»resistance«,»fight«,»dispute«,»demonstration«,»log-offs«, »legal«,»litigation«,»court«,»labour,»trade union«. In addition, we also used context-sensitive keywords, such as »Rappitenderos«, the term for workers at the Colombian delivery company Rappi. 4 In total, 527 incidents of labour unrest in 36 countries were identified for the period 1 January 2017 to 20 May 2020. 5 In other words, labour unrest has been reported by press and media in about 38 per cent of the countries in which the food-delivery platforms are active. Information was collected about the date and location of the incident, the type of action, the number of participants, duration of the action, the collective actors involved, and the cause of the unrest – 2 In some cases, direct information about revenues was available in company reports. Wherever this information was lacking, we relied on owler.com, a website that takes a crowdsourcing approach to estimating the expected revenues of companies in the tech industry. 3 The figures relate to the period under investigation(2017–2020), meaning that they include all countries that the companies were active in during this period, even though they might have withdrawn from or expanded into some other countries in the meantime. 4 Other examples are app-based worker, bike couriers or food-delivery workers. 5 The representation of incidents through GDELT might not capture events that have not been reported, as we rely on the reporting of events in news media. Further data-mining through social media e.g. will help completing the picture. However, relying on news media sources, the list of incidents might be quite complete, as we did cross-check with other sources. 4

THE EXTENT OF LABOUR UNREST Table 2 Data and coding Category Date Location Type of action Dimensions Month and year City and country Strikes, logoffs, demonstrations, legal action, formalisation Notes While a strike is a collective legal withdrawal of labour, a log-off is unofficial or a wildcat form of labour withdrawal. Formalisation captures worker protest leading to institution building such as the formation of works councils, or the negotiating of collective agreements between workers and platform companies which we interpret as use of institutional power. It also captures formalisation of capacity building by increasing union membership or establishing a union which we interpret as signs of associational power. We have recorded all incidents with the intent of formalisation. We cannot assess if the outcome was successful or not. Number of participants Duration of action Actors Cause Gender Wider political and ­social support Source: own elaboration. <10; 11–49; 50–99; 100–499; 500–999; 1000 24h, 2d, 3d, 4d, 5d, 1–2w, 3–8w,>8w Informal group of workers; workers’ collectives (identified by having a distinct name, a Facebook or webpage and not being a union); grass-root ­unions; mainstream trade unions; law firms Pay; employment status; working hours; health and safety; other working conditions; deactivation; union representation; other regulatory issues. Any mentioning of female leaders, activists, or ­female workforce Any mentioning of party support, government, public figures, or NGOs By»legal action«, we mean reference to a particular legal case, which in most reports meant either the beginning of a trial, or the announcement of the intention of a group of workers or law firm to take action. We did not record the duration of any legal action. Consequently, the date recorded for legal actions is intended to capture the beginning of a court hearing of a lawsuit. We interpret this as use of institutional power. Differentiating between union organisations is not a simple matter, as their legal status varies across different countries. Mainstream unions were therefore taken to mean those bodies that were formally recognised as part of a national trade union association and that were typically longstanding. see Table 2 for a more detailed overview. All incidents were coded manually. The»actor« category sheds light on the formation of associational power, allowing us to investigate whether different types of actors pursue different logics of action. The»type of action« category furthermore informs us about the power resources that are used and aimed at in workers’ actions. 4  THE EXTENT OF LABOUR UNREST As noted above, 527 incidents of labour unrest were captured in the dataset for the period between 1 January 2017 and 20 May 2020. The dataset captured incidents in 36 different countries. For the sake of presentation, incidents were grouped across sub-regions, based on the UN geoscheme subregions 6 and then aggregated into continents. The largest number of incidents occurred across Europe (50.8%), followed by Asia, accounting for around a quarter of all incidents(25.3%), and South America, accounting for one-sixth of cases. Far fewer cases were drawn from North America and Australia. No incidents of labour unrest were found in the two African companies. At a national level, the highest number of incidents occurred in China(20.2%), followed by the UK(13.1%), Spain(12.7%) and Argentina (11.4%). Figure 1 shows the timeline of events, logged on a quarterly basis from quarter 1 of 2017 to quarter 2 of 2020. The timeline shows a gradual increase in protests from early 2017 to a significant spike in the third quarter of 2019, with a notable decline since then. It should be noted that the frequencies listed here may include some under-counting, as some protests took place in companies not included in our sample. 6 The United Nations geoscheme is a system which categorises the countries of the world into regional and sub-regional groups. It was devised by the United Nations Statistics Division(UNSD). 5

FRIEDRICH-EBERT-STIFTUNG – Global labour unrest on platforms Figure 1 Occurrence of events over time(in quarters) per continent 100 80 60 40 20 0 frequenzy Q2 Q1 2017 Q 2 3 017 Q 2 4 017 Q 2 1 017 Q 2 2 018 Q 2 3 018 Q 2 4 018 Q 2 1 018 Q 2 2 019 Q 2 3 019 Q 2 4 019 Q 2 1 019 Q 2 2 020 2020 Continent: Source: Leeds Index. Date – Quarter Europe Asia South America North America Australia Of the 19 companies included in the dataset, labour unrest was associated with five in particular. The most notable was Deliveroo, which accounted for more than a quarter of all incidents(28.5%). For the most part, these incidents took place in Europe, specifically in the UK, Spain, France, Belgium and the Netherlands. Labour unrest associated with Meituan accounted for a little over one in ten events(11%) and was exclusively limited to China. The remaining three companies included two global players, Glovo(8.3%), with cases predominantly in Spain, and Uber Eats(6.6%), with incidents located internationally, but with a spike in the UK, and Ele.Me(6.6%), located in China. For the most part, labour actions involved relatively few participants and were of short duration. Eight out of every ten actions involved less than 100 workers and six in ten less than 50. Only a very small proportion involved more than 1,000(2.8%), although around one-sixth of cases (16%) did involve between 100 and 1,000 workers. The vast majority of events(85.1%) lasted less than 24 hours. A little over one in ten cases(11.9%) lasted from one to three days, with very few lasting longer than this. The relatively low frequency of long actions means it was difficult to tease out any association between the duration of actions and the number involved. Labour unrest was not just centred on a single platform company. About a quarter of reported incidents(24.4%) addressed multiple companies, a quarter of the incidents (24.8%) had cross-location coordination and one-fifth (20.5%) included workers from multiple companies. This is an interesting result with regard to associational power. It suggests that workers seek cooperation, collaboration and coalition with each other to defend their interests. Figure 2 Map of protest of Deliveroo, Meituan, Glovo, Uber Eats, Ele.Me Source: Leeds Index. 6

POWER RESOURCES: ASSOCIATIONAL POWER AS THE MAIN SOURCE OF POWER 5  POWER RESOURCES: ASSOCIATIONAL POWER AS THE MAIN SOURCE OF POWER The most obvious insight from our data was the focus of platform workers and their collective action to build associational power. Other sources of power were much less prevalent. We found few signs of societal sources of power; only in very few cases did workers seek support from other social actors. Exceptions were, for example, in the case of Germany, where courier drivers sought support from the Minister for Labour and Social Affairs and who joined a couple of demonstrations. Or, in 2017, the Dutch Labour Party PvdA was actively involved in organising a crowdfunding campaign to finance a court action by a courier from Amsterdam against Deliveroo for bogus self-employment. But, overall, there were few signs of coalitional power. Discursive power was almost impossible to operationalise, but the campaigns we know of from qualitative studies clearly try to mobilise general societal ideas around moral practices. One example, once again from Germany, would be the campaign,»Liefern am Limit«, meaning‘delivering at the limit’ – being at the limit in colloquial German means being at the end, totally exhausted – designed to gather public support against a practice that is seen by most as intolerable and unfair(Trappmann et al. 2020). mainstream. Informal groups of workers were involved in the vast majority of events(85.1%). However, in a quarter of cases each, there were mainstream trade unions(27.6%) and grass-root trade unions(26.9%) involved. In other cases, so-called workers’ collectives(14.6%) were involved. Workers’ collectives like CLAP in France, Liefern am Limit in Germany, KoeriersCollectief/Collectif des coursier.e.s in Belgium or York Courier Community in the UK differ from a loose group of workers, as they have given themselves a name to coordinate their action and forge a collective identity. They represent more formalised bottom-up initiatives, not just a group of workers, but a first step towards more formalisation of associational power. From our dataset, the number of identifiable female activists appeared to be low. Few cases referred to the gender of activists, with only 1.5 per cent of cases making any reference to female leaders, activists or members. The spokespersons, it seems, were mainly men. This does not say anything more generally about women and their activism in the platform economy, but for the sector of food delivery male activists appeared more prominent(compare Churchill 2019; Ford and Hannon 2019). Figure 3 Types of associational power by continent Evidence of institutional power was also rather rare for courier workers. Only 2 per cent of cases of workers’ action related to collective agreements or establishing a works council. In 16.9 per cent of cases, however, workers took legal recourse to defend their interests. This was mainly the case in North America and Australia, where legal action was the most prevalent form of action. This is most likely in the absence of any other form of collective rights for self-employed couriers. One-tenth of legal cases were initiated by law firms, with legal cases frequently initiated by individuals or groups of workers, and some by unions. In North America and Australia, in contrast, the lead actor was mainly mainstream unions, with no grass-root unions involved, and which sought to rely on the law to defend workers’ interest. Although they were not widespread, legal cases as a source of institutional power have led to positive outcomes for platform workers. For example, in March 2019 a group of Rappi couriers in Argentina organised in a union for digital app workers were blocked from getting orders immediately after a meeting with company representatives. The couriers filed a lawsuit and the Argentinian Labour Court ordered the delivery platform to reinstate the workers and cease its anti-union behaviour. While in Spain in November 2018, a Valencian courier was the first Spanish worker to have a court rule against Deliveroo on employment status, with the platform worker being recognised as an employee and awarded compensation. However, those court decisions did not result in a binding regulatory arrangement. Looking more closely at associational power, in the case of platform workers, we see that power seems to stem from workers’ direct voice, not necessarily mediated or represented through membership in unions, be they grass-root or 100 80 per cent (%) 60 40 20 0 Europe Asia North South Australia America America Continent Grassroot union Workers’ collective Mainstream union Group of workers Source: Leeds Index. As Figure 3 illustrates, there were notable regional differences regarding type of associational power. Informal groups of workers were the main actors behind labour unrest in Asia and South America, accounting for 97.7 per cent and 89.4 per cent, respectively. Informal group of workers were slightly less prominent – though still ubiquitous- in Europe (83%), and much less involved in North America(47.4%) 7

FRIEDRICH-EBERT-STIFTUNG – Global labour unrest on platforms and Australia(33.3%). Union involvement was widespread, with unions involved in around half of cases, but with large regional differences. Mainstream trade unions were the dominant actor in Australia(68.8%), relatively active in Europe(37%), and involved in around a quarter of cases (27%) in South America. They were not particularly prevalent in North America(16.7%), and virtually absent in Asia (a mere 1%). Grass-root trade unions were most apparent in South America(42.3%) and Europe(32.9%), less likely in Asia(10%) and absent in North America and Australia. Workers’ collectives were most common in Europe(25.1%), with relatively low reported levels of involvement in South America(10.1%), and Australia(5.9%). They were absent in North America and Asia. In order to get a better understanding of the role and interplay of the different actors, Table 3 presents frequencies of collaboration between actors. To reduce sample bias, China was excluded, as the China Labour Bulletin database does not provide information on the actors involved, and China dominates results for Asia. Informal group of workers were the most prevalent(29.6%), followed by mainstream trade unions and informal group of workers and grass-root trade unions and informal groups of workers, each of which accounted for 14.5 per cent of cases. By contrast, grass-root unions alone were involved in only 5.1 per cent of events, a similar proportion to events that involved only mainstream unions(6.6%). In 5.9 per cent of events, all actors were involved. Table 3 Frequencies of actors and collaboration involving them Actors and collaboration Informal group of workers only Grass-root union& informal group of workers Mainstream union& informal group of workers Workers collective& informal group of workers Mainstream unions only Grass-root, mainstream union, workers ­collective and informal group of workers Grass-root unions only Mainstream, workers collective and informal group of workers Grass-root unions, workers collective and ­informal group of workers Grass-root, mainstream union and informal group of workers Frequency 116 57 57 28 27 23 Percentage 29.6% 14.5% 14.5% 7.1% 6.9% 5.9% 21 5.3% 18 4.6% 8 2.0% 3 0.9% Source: Leeds Index. Note: All coalitions with a 0 frequency are omitted from the table. We also examined specific collaboration – the most frequent collaborations of grassroot and mainstream unions with workers- across the different regions. Figure 4 shows that both grass-root and mainstream unions had a strong presence in Europe. In particular, in Europe, actions with a coalition between grass-root unions and an informal group of workers took place in 14.6 per cent of cases and with coalitions between mainstream unions and an informal group of workers actions in 15.4 per cent of cases(for the rest of the actor constellations and their unrest cases compare the ANNEX). Grass-root and mainstream unions were also involved in a large number of cases in South America, with coalitions between grass-root unions and an informal group of workers reaching 20.2 per cent, compared to coalitions between mainstream unions and informal group of workers in just 10.1 per cent of cases. In the US and Australia, groups of workers only collaborate with mainstream unions in 15.8 per cent of cases for the US and 23.5 per cent for Australia. In Asia, unions play a minor role, but where they are involved they tend to be grass-root unions. Figure 4 Actors and collaboration by continent 25 20 per cent (%) 15 10 5 0 Europe Asia North South Australia America America Continent Grassroot union and Group of workers Mainstream union and Group of workers Source: Leeds Index. Clearly, then, we see a strong role for self-organisation of platform workers. It was predominantly the workers who took direct action. However, in almost half of all cases this was with the support of trade unions. Where workers were supported by trade unions, there was an equal level of representation amongst both mainstream and grass-root unions, though with notable regional variations. So, while this is an emerging sector of the economy, and mainstream unions have been seen to be reluctant(Gumbrell-McCormick, 2011) to organise outsiders such as precarious platform workers, this was not immediately apparent in our data, which shows they were as equally active as grass-root trade unions. Interestingly enough, there were just three cases in our dataset where grass-root and mainstream union were involved together. This begs the question whether this was due the fact that grass-root and mainstream unions were pursuing difference goals? This hypothesis is first explored by examining the types of issues precipitating labour unrest. 8

WHAT TYPE OF ACTION AND FOR WHAT GRIEVANCE? 6  WHAT TYPE OF ACTION AND FOR WHAT GRIEVANCE? of legal arbitration, and in the US given the weakness of associational power of workers. The most frequent types of action(see figure 5) were strikes and log-offs,(40.4%), followed by demonstrations(34.2%), legal action(16.3%), and formalisation(3.7%, comprised of 2% formalisation through institution-building and 1.7% formalisation through capacity-building). Figure 5 Type of action by region 100 80 per cent (%) 60 40 20 0 Europe Asia North South Australia America America Continent Demonstrations Institutionalisation Legal action Source: Leeds Index. Other action Strikes/Log-offs At a regional level, strikes and log-offs were most frequent in Asia, accounting for 64.7 per cent of actions. By contrast, such labour unrest was less frequent in South America (25.8%), Europe(38.6%) North America(5.3%), and Australia, where no such events were recorded. Demonstrations were most frequent in South America(53.9%), followed by Europe(32.6%), Asia(28.6%), North America(5.3%) and Australia(23.5%). Nearly two-thirds of events in North America(63.2%) involved legal action, a form of event that was nearly as common in Australia(58.8%), but far less apparent across Europe(18.7%). Strikes and log-offs were also associated with smaller groups of workers, less than 50, compared to demonstrations, which were associated with slightly larger groups of workers, between 100 and 499. The type of action dominant in each region might be related to the prevailing employment relations regime. It is quite likely that in countries like China workers relied more on log-offs and disruptive actions in the absence of an open media system that would allow any impact in the wake of demonstrations. Legal action might be more likely in countries like Australia because of its historical tradition Table 4 presents a list of the grievances that motivated platform workers’ actions. While the findings are presented in relation to the frequency of each action, it is important to note that it was relatively common for actions to be chronicled in terms of a number of motivating factors. Distributive concerns were by far the most common reason for labour unrest, with a little under two-thirds of our dataset(63.4%) related to actions involving pay. Three other issues were the motivating factor in around a fifth of cases, employment status(22.3%), working conditions(20.2%) and health and safety(17.1%). Other issues occurred far less frequently – in no more than ten per cent of cases for each. Of the less frequently reported drivers of labour unrest, deactivation was reported in just 6.5 per cent of cases and a demand for union representation even less than this(4.4%). Given the focus of attention on issues pertaining to employment status and algorithmic management as key signifiers of platform work, it is intriguing to find that the key grievance amongst platform workers – across most regions – was the traditional distributive issue of payment. This is the same grievance traditionally dominating the issues of concern to non-platform workers in the non-digital world of work. Apparently, the concerns of platform workers were far less different than those of traditional workers and that the current literature would have us believe. Table 4 Grievances prompting labour unrest(rank order:%)(N=525) Actors and collaboration Pay Employment status Working conditions Health and safety Deactivation Other issues Non-pay benefits Working hours Union representation Other regulatory issues Percentage 63.4% 22.3% 20.2% 17.1% 6.5% 6.3% 5.0% 4.8% 4.4% 3.2% Source: Leeds Index. Note: the data do not add up to 100, as multiple drivers could be recorded per event. 9

FRIEDRICH-EBERT-STIFTUNG – Global labour unrest on platforms per cent (%) Figure 6 Grievances by region 100 80 60 40 20 7  CONCLUSION Our findings suggest a remarkable amount of labour unrest involving workers on food delivery platforms. These incidents of labour unrest span many different regions of the world, and our sample covered labour unrest in 36 countries. We take this as a sign of rising associational power among platform workers. Associational power is clearly worker-led, with most incidents of unrest stemming from workers themselves. Despite the conditions of platform work, and the labour process in particular that have been described as detrimental to mobilising and organising, courier workers were seen to take action to try to improve their working lives. There were many cases that span across multiple companies and multiple locations, suggesting platform workers were developing the capacity to wage concerted campaigns and build solidarity. 0 Europe Asia North South Australia America America Continent Pay Working hours Working conditions Employment status Union representation Source: Leeds Index. Other regulatory issues Deactivation Health& Safety Non-pay benefits Again, we find notable variations according to region. A low level of labour unrest around deactivation and union representation was the same everywhere. Pay was the most dominant driver of labour unrest across all regions, with the exception of the US, where employment status was the most significant reason for unrest(see figure 6). To some extent this was a surprising finding. We might have expected struggles over employment status to be more dominant in those countries with strong employment rights, rather than in countries such as the USA. That said, the absence of wider employment rights in the USA, coupled with a widespread propensity for litigation, has meant that struggles over employment status have been a defining feature of platform labour unrest in key American states. In South America, health and safety was dominant, and this appears to have increased during the COVID19 pandemic. In the wake of the COVID19 crisis, and the discursive classification of food delivery workers as»essential workers« exposed to higher risk, couriers have demanded higher pay, provisions for healthcare, and have frequently criticised the insufficient protective gear provided by platform companies(mainly Pedidosya, Glovo, SinDelantal and Uber Eats). These strikes have been sector-wide across countries, representing»the first real example of an international, sector-wide, strike movement in the gig economy.«(Howson et al. 2020). The pandemic provided the impetus and platform for workers to raise their voices against underlying structural injustices of their platform work(Howson et al. 2020). Grass-root-driven unrest occurred in many cases, however, with help or support or in coordination with trade unions. Trade unions were involved in almost half of the incidents. While both types of unions- grass-root and mainstream appeared to be important, although varyingly in this respect depending on the country, there appeared to be little evidence of cooperation. In some cases, worker-led unrest led to unionisation, both increasing membership of existing unions or establishing a new union. 7 We saw only a small number of workers’ collectives active in platform labour protest, which offers little by way of more generalised support for early studies that suggest workers’ collectives can act as important bridges between grass-root protest and established unions(Jesnes et al., forthcoming). Associational power varied widely across different regions. The most diverse set of actors were involved in Europe. Here, the evidence suggests that associational power was rising, built from the bottom-up with a higher number of workers’ collectives emerging(25.1%), 32.9 per cent grassroot unions active, 37.0 per cent mainstream unions active. A variety of actors were also involved in South America: workers(89.4%); both types of unions(grassroots 42.3%; mainstream 27%); and a rising number of workers’ collectives(10.1%). In Asia, it was mostly workers(97.7%) who acted, without much support from unions(1% mainstream; 10% grass-root). In Australia, mainstream unions were particularly active(68.8%), with no grass-root unions and only a small share of worker-led unrest(33.3%). Associational power appeared to be least developed in the US. In this 7 In September 2018, CGT set up a special branch of the union for Glovo’s couriers in Granada, Spain. In February 2020, the Ontario Labour Board in Canada ruled that Foodora couriers have the right to form a union under the Umbrella of the Canadian Union of Postal Workers. The majority of couriers in Toronto and Mississauga then voted to join the union, but Foodora withdrew from the Canadian market shortly afterwards. In October 2019, 17 Uber Eats couriers in Tokyo established a union to negotiate on working conditions and to demand employee status. In the same month, workers from Rappi, Glovo, Uber and Cabify registered the Asociación de Personal de Plataformas(Platforms Staff Association), the first union for digital platforms in Argentina. 10

case, it was mainly institutional power via legal action that was used to defend workers’ interests. Looking more closely at the role of different actors, we did not find that grass-root unions and self-organising workers pursued different goals, or acted very differently than more mainstream unions. Vandaele’s(2018) assertion, although only based on western Europe, that grass-root and mainstream unions pursue different logics, did not appear to be supported by the global data. Workers and trade unions alike drew on institutional power sources or sought to increase institutional power, but only in a minority of cases. At the company and sectoral level- via works councils and collective agreements- our data show it only happened in three countries, while at the macro level reliance on the law – by taking employers to court – was concentrated in the USA and Australia. Establishment of societal support also only emerged in a negligible number of cases. The most dominant issue for labour unrest was pay. This could be viewed as a surprising result, given the emphasis on employment status and algorithmic control as sources of contention in the existing literature. Disputes over pay reflect classic themes of conflict over the level of exploitation and the level of compensation for labour and as such constitute the main driver of protest. This is again at the core of established trade unions’ self-understanding of what to fight for. Therefore our data suggest a solid basis for trade unions to organise platform workers in the future and an important step towards a better understanding of the global dynamics of this movement. Conclusion 11

FRIEDRICH-EBERT-STIFTUNG – Global labour unrest on platforms REFERENCES Alimahomed-Wilson, J.,& Ness, I.(Eds.).(2018) Choke points: Logistics workers disrupting the global supply chain. Pluto Press Anwar, M. and Graham, M.(2019) Hidden transcripts of the gig economy: labour agency and the new art of resistance among African gig workers. Environment and Planning A: Economy and Space, 52(7), pp. 1269–1291. Atzeni, M., Ness, I.(2018)(Eds.) Global Perspectives on Workers’ and Labour Organizations. Springer. Bank Muñoz, C.(2017) Building Power From Below: Chilean Workers Take on Walmart. Ithaca, NY: Cornell University Press. Bordogna, Lorenzo and Gian Primo Cella(2002) Decline or transformation? Change in industrial conflict and its challenges. in Transfer 04: 585–607. Brinkmann, U.,Choi, H.-L,Detje, R.,Dörre, K.,Holst, H.,Karakayali, S. and Schmalstieg, C.(2008): Strategic Unionism: Aus der Krise zur Erneuerung? Umrisse eines Forschungsprogramms. Wiesbaden. Briziarelli, M.(2018) Spatial politics in the digital realm: the logistics/precarity dialectics and Deliveroo’s tertiary space struggles. Cultural Studies, 33(5), pp. 823–840. Cant, C.(2019) Riding for Deliveroo: Resistance in the new econo my. Cambridge, UK; Medford, MA: Polity. Cant, C. and Woodcock, J., 2020. Fast Food Shutdown: From disorganisation to action in the service sector. Capital& Class. Chesta R.E., Zamponi, L. and Caciagli C.(2019) Labour Activism and Social Movement Unionism in the Gig Economy. Food Delivery Workers Struggles in Italy. Partecipazione e Conflitto, 12(3), pp. 819–844. Chun, J.(2009) Organising at the Margins: The Symbolic Politics of Labor in South Korea and the United States. Ithaca: Cornell University Press. Churchill, B., Craig, L.(2019) Gender in the gig economy: Men and women using digital platforms to secure work in Australia. Journal of Sociology, 55(4), 741–761. Doellgast, V., Lillie N. and Pulignano V.(2018) Reconstructing solidarity: Labour unions, precarious work, and the politics of institutional change in Europe. Oxford, UK: Oxford University Press. Eaton, A., Schurman, S., and Chen, M.(Eds.)(2017) Informal Workers and Collective Action: A Global Perspective. ITHACA; LONDON: Cornell University Press. Ford, M., and Honan, V.(2019) The limits of mutual aid: Emerging forms of collectivity among app-based transport workers in Indonesia. Journal of Industrial Relations, 61(4), 528–548. Fox-Hodess, K.(2019) Worker Power, Trade Union Strategy, and International Connections: Dockworker Unionism in Colombia and Chile. Latin American Politics and Society, 61(3), 29–54. Gumbrell-McCormick, R,(2011) European Trade Unions and Atypical Workers. Industrial Relations Journal, 42(3), 293–310. Hinz, S.(2018) The Ambivalence of Structural Power: Alternative Trade Unions Challenging Transnational Automotive Companies in Russia. Global Labour Journal, 9(2), 167–181. Howson, et al.(2020) Just because you don’t see your boss, doesn’t mean you don’t have a boss: Covid-19 and Gig Worker Strikes across Latin America. International Union Rights. Volume(27)3: 20–21. Jesnes, K., Neumann, D., Trappmann, V. and de Becdelièvre, P. (2021) Food delivery couriers’ protests in France, Germany and Norway: all the same or different? The role of workers’ collectives. In Drahokoupil, J., K. Vandaele(Eds.) A Modern Guide to Labour and the Platform Economy. Edward Elgar(in press) Kirk, E.(2020) Contesting»bogus self-employment« via legal mobilisation: The case of foster care workers. Capital& Class. First published online: 17 February. https://doi.org/10.1177/0309816820906355 Lévesque, C. and G. Murray(2010) Understanding Union Power: Resources and Capabilities for Renewing Union Capacity. European Review of Labour and Research, 16(3): 333–50. Moody, K.(2017) On New Terrain: How Capital is Reshaping the Battleground of Class War. Haymarket Books. Murray, G.(2017) Union Renewal: What Can We Learn from Three Decades of Research? Transfer: European Review of Labour and Research, 23(1): 9–29. Minter, K.(2017). Negotiating labour standards in the gig economy: Airtasker and Unions New South Wales. The Economic and Labour Relations Review, 28(3), 438–454. Offe, C., Wiesenthal, H.(1980) Two Logics of Collective Action: Theoretical Notes on Social Class and Organizational Form. Political Power and Social Theory, 1: 67–115. Rizzo, M., Atzeni, M.(2020) Workers’ Power in Resisting Precarity: Comparing Transport Workers in Buenos Aires and Dar es Salaam. Work, Employment and Society. Schmalz, S.,Dörre, K.(Eds.)(2013): Comeback der Gewerkschaften?: Machtressourcen, innovative Praktiken, internationale Perspektiven. Frankfurt a. M. Schmalz, S., Dörre, Klaus(2014): Der Machtressourcenansatz: Ein Instrument zur Analyse gewerkschaftlichen Handlungsvermögens. Industrielle Beziehungen, 21(3), pp. 217–237. Schmalz, S; Ludwig, C and Webster, E.(2018) The power resources approach: developments and challenges Global Labor Journal 9(2) 113– 134 Silver, B. J.(2003). Forces of labor: workers’ movements and globalization since 1870. Cambridge University Press. Tassinari, A., and Maccarrone, V.(2017) The mobilisation of gig economy couriers in Italy: some lessons for the trade union movement. Transfer: European Review of Labour and Research, 23(3), 353–357. 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), 35–54. Trappmann V., Neumann, D., Stuart, M.(2020) New forms of labour unrest in platform work: Insights from Germany. Manuscript Vandaele, K.(2018). Will trade unions survive in the platform economy? Emerging patterns of platform workers’ collective voice and representation in Europe. Emerging Patterns of Platform Workers’ Collective Voice and Representation in Europe(19 June 2018). ETUI Research Paper-Working Paper. Vandaele, K., Piasna, A.,& Drahokoupil, J.(2019). ›Algorithm breakers‹ are not a different ›species‹: attitudes towards trade unions of Deliveroo riders in Belgium. Are not a Different ›Species‹: Attitudes Towards Trade Unions of Deliveroo Riders in Belgium(12 June 2019). ETUI Research Paper-Working Paper. Von Holdt, K.,& Webster, E.(2008) Organising on the periphery: new sources of power in the South African workplace. Employee Relations, 30(4), 333–354. Webster, E.(2015). Labour after globalisation: Old and new sources of power. Labour and transnational action in times of crisis, 115–27. Wood, A. J., Lehdonvirta, V.,& Graham, M.(2018). Workers of the Internet unite? Online freelancer organisation among remote gig economy workers in six Asian and African countries. New Technology, Work and Employment, 33(2), 95–112. Wright, E. O.(2000). Working-class power, capitalist-class interests, and class compromise. American Journal of Sociology, 105(4), 957–1002. Zajak, S.(2017). International allies, institutional layering and power in the making of labour in Bangladesh. Development and Change, 48(5), 1007–1030. 12

Annex ANNEX Actors& Collaboration Informal group of workers only Grassroot unions only Mainstream unions only Workers collectives only Grassroot unions& Informal group of workers Mainstream unions& informal group of workers Informal group of workers& workers collectives Grassroot TU, Mainstream TU, Informal group of workers Mainstream TU, Informal group of workers& workers collective Grassroot TU, Informal group of workers& workers collective All coalitions (Grassroot, Mainstream TU, ­ workers collective& informal group of workers) All N% 235 44.7 22 4.2 26 4.9 0 65 12.4 58 11.0 28 5.3 3 0.6 18 3.4 8 1.5 23 4.4 Europe N% 72 27.0 16 6.0 17 6.4 0 39 14.6 Asia N% 118 88.7 2 1.5 0 0 8 6.0 41 15.4 1 0.8 25 9.4 0 1 0.4 0 17 6.4 0 8 3.0 0 17 6.4 0 US N% 6 31.6 0 0.00 0 0 0 3 15.8 0 0 0 0 0 SA N% 38 42.7 4 4.5 3 3.4 0 18 20.2 9 10.1 3 2 2.2 0 0 6 6.7 Africa N% 1 100 0 0 0 0 0 0 0 0 0 0 Australia N% 0 0 0 0 6 35.3 0 0 4 23.5 0 0 1 5.9 0 0 13

IMPRINT ABOUT THE AUTHORS Vera Trappmann is an Associate Professor in the Work and Employment at Leeds University Business School and member of the Centre of Employment Relations Innovation and Change(CERIC). Her research is on precarisation of work, organisational restructuring and labour’s response. Ioulia Bessa is a University Academic Fellow, based in the Centre of Employment Relations Innovation and Change (CERIC). Her research focus includes non-standard forms of employment and their connection to precarity, underemployment/unemployment. She has experience of working with a range of secondary datasets(national and international). Simon Joyce is a Research Fellow at the Digital Futures of Work(Digit) Research Centre, at Leeds University Business School. He has been researching platform work since 2016. Friedrich-Ebert-Stiftung| Global Policy and Development Hiroshimastr. 28| 10785 Berlin| Germany Responsible: Mirko Herberg| International Trade Union Policy Phone:+49-30-269-35-7458| Fax:+49-30-269-35-9255 www.fes.de/gewerkschaften To order publications: Julia.Schultz@fes.de Commercial use of all media published by the FriedrichEbert-Stiftung(FES) is not permitted without the written consent of the FES. Denis Neumann is a Researcher at Centre of Employment Relations Innovation and Change(CERIC). His research interests are power relations in platform work and young precarious workers relations with trade unions. Mark Stuart is the Montague Burton Professor of Human Resource Management and Employment Relations at Leeds University Business School, University of Leeds. He is the co-director of the ESRC Digital Futures at Work Research centre. Charles Umney is an Associate Professor in the Work and Employment Relations Division at Leeds University Business School. He has various research interests including platform worker protest; the social effects of marketising public sector reforms; and working life in arts and culture. The views expressed in this publication are not necessarily those of the Friedrich-Ebert-Stiftung(or of the organization for which the author works). This publication is printed on paper from sustainable forestry. ISBN 978-3-96250-709-1

GLOBAL LABOUR UNREST ON PLATFORMS The case of food delivery workers Labour unrest by platform workers is an important phenomenon in the new world of work. This study examines patterns of platform labour unrest on a global scale, drawing from a database of over 500 instances of labour unrest in the food delivery sector. Results show that labour unrest has been growing in recent years and has spread across a large number of countries around the world. Research findings show that labour unrest is driven by workers themselves, but increasingly receives support from trade unions. Relying hitherto mainly on the ability to act collectively(associational power) and to mobilie public support for their cause(societal power), trade unions through their political and legal experiences add a further dimension to the power resources available. Issues driving unrest are not necessarily linked to novel characteristics of platform work. In most cases and regions, they have their source in»classic« conflicts of distributional quality like pay and working conditions. Employment relations and health protection under the Covid-19 pandemic add, however, to the urgency of finding solutions to the causes of workers‘ unrest. For further information on this topic: https://www.fes.de/lnk/transform