“The Food Delivered is More Valuable Than My Life”: Understanding the Platform Precarity of Online Food-Delivery Work in China

Abstract

This article investigates the precarious labour conditions of Chinese food-delivery drivers in the platform economy. Drawing on one year of ethnographic fieldwork where the author worked as a food-delivery driver in Shanghai, the three key forces producing precarity in the platform labour regime are explored: (i) the platform circumvents its employer responsibilities for drivers by outsourcing the labour services of food delivery to third-party labour-hires companies; (ii) predatory algorithmic management is leveraged by the platform to control the labour process for excessive exploitation; and (iii) the institutional deprivation of citizenship rights of the rural migrants converts drivers into urban denizens with a vulnerable socio-economic labour environment. These determinants combine to produce low-paid, insecure, uncertain, and dangerous working conditions which food delivery drivers have limited power to resist both at individual and collective levels. Building on these findings, this article argues that the peculiar intersection of bogus triangular employment relations, predatory algorithmic control, and the subservient citizenship of rural migrants, produces precarity in the platform labour regime. The article highlights the role of the state and management in producing the precarity experienced by Chinese food-delivery drivers and contributes to understanding the work precarity of the platform economy in the digital age.

Key Words:

• Digital economy
• food delivery
• migrant workers
• platform work
• platform economy
• precarious work

A recent focus in the sociology of work has been on the association between platform-mediated employment, precarious labour conditions, and especially the digital repackaging of low-skilled “gig” work (see, for example, Kahancová, Meszmann, and Sedláková 2020; Schor et al. 2020). Extant scholarship indicates that platform-based gig work is one of the “accelerants of precarity” (Vallas and Schor 2020, 279), which results in further labour de-standardisation, commodification and casualisation (Standing 2016; Kalleberg and Vallas 2017; Zwick 2018). Platform-based food-delivery work, which is an important sector of the platform economy, has produced what may be considered ultra-precarious labour conditions for workers (see, for example, Cant 2019; Goods, Veen, and Barratt 2019; Tassinari and Maccarrone 2020; Huang 2021). In China, there are about eight million rural migrant workers making a living and crafting their lives in the booming online food-delivery industry (Lei 2021; Huang 2022).

This article investigates the precarious lives of food-delivery drivers in China. It argues that an explanation of how their “precariousness” is produced and reproduced requires an understanding of the role of the state, especially with respect to how the state denies rural migrant workers citizenships rights, as well as the role of management using subcontracting and algorithmic technology that shapes the labour process. These determinants combine to produce low-paid, insecure, uncertain, and dangerous working conditions which food-delivery drivers have limited individual and collective power to resist.

The article is structured as follows. First, the following section offers a brief review of the concept of precarity and literature regarding how precarity is produced in the platform labour regime. Second, some comments are made about the context of the Chinese operation of food-delivery work and methodology used to collect and analyse data. Third, the main findings are presented in four sections with a discussion of the bogus triangular employment relations, predatory algorithmic control, subservient citizenship of drivers and their suppressed labour resistance. Finally, a conclusion summarises key arguments made.

Precarity and the Platform Labour Regime: The Literature

The concept of precarity has been widely used as an analytical tool to examine the labour conditions of non-standard work arrangements. Since the neo-liberal liberation of industrial relations that erodes Fordist employment regimes, this concept has been developed by a wide range of scholars. For example, Rodgers (1989, 3) argues that precarity is characterised by employment insecurity, lack of state protection for workers against unfair treatment by management, as well as a loss of control over the labour process. This perspective highlights the issue of control of the labour process in defining precarity. Adopting a wider perspective, Standing (2011) views precarity in light of the complex relations between labour, capital and state, arguing that precarious work should be understood in terms of three dimensions: distinctive relations of production (flexible contractual relations), distinctive relations of distribution (money wages without non-wage benefits), and distinctive relations to states (denizens who are not entitled to whole citizenship rights). More recently, Kalleberg and Vallas (2017, 1) define precarious work as “work that is uncertain, unstable, and insecure and in which employees bear the risks of work (as opposed to businesses or the government) and receive limited social benefits and statutory protections.” Taken together these differing definitions assist in an understanding of the characteristics, causes and impact of precarious work. Put simply, precarious work is associated with unregulated and casual employment, on-demand, outsourced labour which is often organised by labour recruitment agencies. Precarious workers experience low and unstable incomes, work insecurity, and possess limited individual and collective power.

In the digital age, alongside the rapid proliferation of work platformisation, a new digital regime of work is emerging – the platform labour regime. This workplace regime generates deeper levels of precarity than that associated with traditional insecure work (Standing 2016; Burrell and Fourcade 2021, 218–220). The concept of platform labour regime refers to a type of labour process whereby a platform company employs a digital platform to mediate the labour exchanged between workers and the company’s clients or customers (Howcroft and Bergvall-Kåreborn 2019, 23–25; Gandini 2019, 1040). Under this regime, the digital platform uses advanced algorithms to organise and control the labour process. Workers are viewed as self-employed independent contractors, reliant on low-quality, low-skilled and low-paid “gig” or irregular work to survive (Vallas and Schor 2020, 273; Kahancová, Meszmann, and Sedláková 2020). Studies show that a majority of platform workers are drawn from marginalised social groups such as ethnic minorities, females, students or migrants (see Alberti et al. 2018). Veen, Barratt, and Goods (2020, 390) claim that platform work can be viewed as a digital repackaging of traditional informal work that extends and intensifies precarity.

Studies have focused on the ways in which precarity is produced and reproduced within platform labour regimes. One strand of the literature has examined the role of management’s use of digital platforms to generate precarity. Platforms create new forms of organisation of labour that disadvantage workers. Kalleberg and Vallas (2017) argue that companies use digital platforms as their organisational structure to shift economic risks to workers by redefining workers as independent contractors. As most platform workers are often paid by piece-rate and not allowed to access work-related benefits and social welfare, they necessarily face economic instability and insecurity (Schor et al. 2020; Vallas and Schor 2020, 279–280). In addition, platforms build virtual organisations to organise workers, which facilitates the physical atomisation of work and fragmentation of labour. Therefore, the traditional organisational scaffold is diminished as workers are virtually organised and are geographically scattered and spatially dispersed (Kaine and Josserand 2019, 495). This impedes the capacity of workers to build and sustain conventional forms of solidarity and collective power for interest participation and representation, such as trade unions (Tassinari and Maccarrone 2020; Cant 2019).

Platform labour regimes use algorithmic technologies to organise the labour process and define the roles and functions of management. Huang (2022) argues that management’s use of algorithmic technologies enables a high degree of work automation and atomisation which accelerates the deskilling and degradation of work. Moreover, management’s deployment of digital technologies provides them with a range of techniques to control and exploit labour (Duggan et al. 2020). For example, information asymmetry is used in the labour process to constrain workers’ autonomy, and algorithmic surveillance is used to closely monitor workers (see Kellogg, Valentine, and Christin 2020; Zuboff 2015). In addition, data-driven game-based tactics (such as rankings, ratings and badges) are used to evaluate workers’ productivity and encourage workers to willingly engage in their own exploitation (see, for example, van Doorn and Chen 2021; Galière 2020). Combined, these various techniques and use of digital technologies exert levels of control that are more stringent than those that exist in Fordist workplaces, intensifying processes of surplus extraction and subjecting workers to a production process over which they have limited power or agency to control.

Another strand of the burgeoning literature on precarity has examined the role of the state in producing and reproducing precarious work under platform labour regimes. It has shown that states have been enthusiastically involved in pursuing neo-liberal socio-economic reforms aimed at securing the deregulation and greater flexibility of employment relationships (Kalleberg 2009, 2; van Doorn 2017, 898). The lack of willingness on the part of states to regulate and limit the use of platform work as well as legislate to protect platform workers is symptomatic of the adoption of neo-liberal policies that fundamentally aim to enhance the power of capital at the expense of labour (Minter 2017, 450). Studies show that many platform workers are drawn from subordinate social groups, especially migrants who do not enjoy full citizenship and associated social protections (Standing 2016, 105). Platforms use these conditions to intensify labour exploitation and to constrain potential organised labour resistance (Glavin, Bierman, and Schieman 2021; Huang 2021).

In this context, the approach of Alberti and colleagues (2018) offers a fruitful approach for investigating precarity. They argue that the state and management combine to produce and reproduce precarious work. They draw a distinction between production-based and citizenship-based precarity. States drive “citizenship precarity” by limiting the access some social groups have to various social protections and workplace-based rights and benefits. Management drives “production-based precarity” through the imposition of risky working conditions and contractual forms such as zero-hours contracts and subcontracting. Management also contributes to this aspect of precarity through strengthened control of the labour process and obscuring the processes of value creation. Alberti and colleagues (2018, 451) also introduce the concept of “implicit precarity” which refers to the kinds of subjective experiences commonly encountered by precarious workers: lack of social recognition, marginalisation, and social isolation. This approach is useful for investigating the Chinese platform labour regime and the impacts this has on labour.

As discussed further below, the precarious nature of work associated with food-delivery platform work has been shaped by both platform companies and their managers as well as central and local political authorities. As part of its aim to rapidly industrialise and urbanise, the Chinese state has pursued a race-to-the-bottom labour policy through the mobilisation of cheap, unskilled, and readily disposable labour drawn from rural areas. The Hukou system (of household registration) effectively marginalises these rural migrants, and deprives them of their citizenship rights, including rights to reside in urban areas, rights to education, medical services, and a range of other social rights and benefits. The precarious nature of their citizenship is reinforced by the platform companies who use third-party staffing and recruitment agencies as well as algorithmic-based systems of control to extract surpluses from food delivery drivers (Lei 2021; Huang 2022). In sum, the approach of Alberti and colleagues allows us to understand and explain different dimensions of precarity and its impact on workers as the product of both state and management action.

Precarity and the Platform Labour Regime: The Chinese Context and Methods

Context

This article is based on one year of ethnographic fieldwork conducted at a popular online food-delivery platform in Shanghai. To guarantee anonymity, the name of the platform is replaced by X-platform. Choosing X-platform and Shanghai for participant observations was deliberate. The online food delivery industry has been growing rapidly in China over the last decade and China has the world’s largest food-delivery market, up to approximately $110 billion in 2020 (SIC 2021). The expansion of the food-delivery industry is monopolised by two platform-based start-ups, which were acquired by digital platform giants Tencent and Alibaba respectively, occupying more than 95% market share in total. At the end of 2020, X-platform has more than half of the market share and “employs” approximately four million registered food delivery drivers (SIC 2021).

The rapid proliferation of the online food delivery industry is a microcosm of China’s changing industrial structure and national policy. China’s economy is experiencing a dramatic transition from a labour-intensive “world factory” to a technology-oriented industrial powerhouse, and China is the world-leading power in developing digital technology and a platform economy (Sigley and Powell 2022). Shanghai has China’s largest population of rural migrant workers from poor provinces. In response to what Roberts and colleagues (2021, 37) identify as the central government’s “industrial digitalisation and manufacturing robotisation” ambitions, Shanghai’s authorities have implemented the “Digital Shanghai” initiative. In this context, the food-delivery industry has become a labour pool for unemployed workers during the ongoing industrial upgrading. Nearly all drivers are young and middle-aged ex-factory workers, ejected from shrinking labour-intensive low-end industrial sectors and forced into the expanding urban surplus population who rely on hyper-precarious gig work to survive.

Data Collection and Analysis

To understand the drivers’ experiences of precarity both in the platform labour regime and in the wider political economy and society, a combination of participant observations and in-depth interviews was adopted. Initially this involved five months of study while working as a driver from September 2019 to February 2020. This experience not only permitted an understanding of the significance of algorithmic architecture for organising work and controlling the labour process but also helped make sense of the broader social-political environment that drivers faced. A daily fieldwork diary was composed to record the everyday reality of drivers’ working experiences. The bulk of data were collected through substantial informal interviews and conversations with drivers, regarding contractual arrangements, techniques of algorithmic control, and drivers’ lived experiences. Due to the sudden outbreak of Covid-19, the research was shifted to virtual ethnography. In this stage, in-depth semi-structured interviews were conducted to supplement already collected material. A total of 58 migrant drivers were interviewed online through WeChat. Based on the social connections of the drivers, the participants were identified through snow-ball sampling. The description of informants is in Table 1.

Table 1. Participant information

Gender	Male	57
	Female	1
Age	Mean	27
	16–29	35
	30–44	21
	Above 45	2
Hukou status (household registration)	Rural	58
	Urban	0
Educational level	Vocational degree and above	3
	Senior secondary school	27
	Junior secondary and below	28
Engagement type	Full-time	45
	Part-time	13
Driver classification	Piece-rate driver	21
	Shift driver	37
Working hours	Below 8	9
	8–12	29
	Above 12	20
Economic dependence	Main income	46
	Supplement income	12
Sending province	Anhui (18), Henan (16), Sichuan (9), Hunan (8), Hubei (7)
Total		58

Data were organised by thematic analysis through inductive and deductive approaches, moving between literature on precarity and the fieldwork data. Whenever possible in-vivo codes and descriptive codes were used drawing from drivers’ descriptions of their lived experiences. To protect the drivers, all information is anonymised. Four key themes with respect to employment relations, algorithmic management, subservient citizenship, and labour resistance were identified, as analysed in the following sections.

Precarity and the Platform Labour Regime: Findings

Bogus Triangular Employment Relations

Recent legislative changes in some countries require platform companies to treat on-demand workers as formal employees (Chan, Nair, and Rhomberg 2019). This has not happened in China. While the nature of the Chinese legal system makes reform difficult, the economic and employment policies and ambitions of the Chinese state exacerbate the difficulties faced by drivers (Lin 2022; Chan, Nair, and Rhomberg 2019; Lei 2021). The state views the digital economy as an industry which has the potential to power the economy and achieve global competitiveness. To promote the growth of the platform economy, the state has provided a flexible regulatory environment for the platforms. The state’s ambition to promote employment more generally has also shaped its interventions into labour relations in the platform economy. Platform companies have used these wider political-economic circumstances to develop strategies to avoid their responsibilities to drivers.

X-platform uses third-party companies to obscure the nature of its contractual relations with its drivers. Claiming to be a technological innovation firm, the company outsources its labour requirement to contractors. Contractors run the business of food delivery and assume responsibility for employing as well as managing drivers. X-platform plays the role of providing technical and managerial support for its contractors. There are two main types of contractors: city-based franchisees and labour service companies. This corresponds to the two different ways that drivers are employed: shift drivers (who work on scheduled time shifts) and piece-rate drivers (who work on random piece-rates). The franchisees manage the work of shift drivers through work stations established in different zones while piece-rate drivers are controlled by labour service companies using digital applications. Through these arrangements, X-platform can evade legal responsibility for drivers. However, the labour process and management of workers’ performance is actually controlled by X-platform via its digital platform.

Based on these arrangements, X-platform is able to further externalise risks and responsibilities to workers through the use of “clickwrap agreements” – a digital prompt that allows individuals to accept or decline a digitally-mediated agreement. This compels drivers to agree to certain terms and conditions before being able to use the application to access work. In the registration process, applicants must click an “agree” button and consent to a clickwrap agreement, and associated terms and conditions. If a driver fails to agree, their registration will be cancelled, and they will be denied access to work through the X-platform application. The clickwrap agreements clearly indicate that X-platform is separate from drivers, forcing drivers to acknowledge that they are independent contractors working for third-party companies. By having to sign these clickwrap agreements, workers are forced to bear the risks of work. The following are excerpts of clickwrap agreements also known as online food delivery service agreements:

Special notes: X-platform is an Internet information platform which provides information matching services between you, labour service companies and customers. You do not have any form of labour/employment relationships with the X-platform company … You and third-party labour services companies are in partnership relations. We are currently providing a free platform for you to match your labour service with customers. However, we reserve the rights to charge you to use our platform … You must bear all the risks and consequences when providing labour service to customers, including traffic accidents, fines incurred by public sectors, disputes with restaurants and customers …

Most franchisees subcontract the labour service of food delivery to other labour service companies. In fact, in China, most of these labour service companies are ghost companies without adequate legal standing and are not subject to any industry regulations (Lin 2022, 437). In these circumstances, most drivers claimed that they were unbale to identity their employer. Whether employed as shift or piece-rate drivers, all drivers complained that they had no option but to accept the unequal terms and conditions contained in agreements in exchange for being granted access to work opportunities. As one explained: “I don’t know who my employer is. None of my fellow drivers do. We understand it’s important to work for a legal company. But it is more important for us to get a job to survive.”

These complex and obscure relationships between X-platform, sub-contractors and drivers are fundamental to the exploitation of migrant drivers. As drivers are treated as self-employed, they are requested to provide their own means of production, such as electric bikes, helmets, food containers and uniforms. Drivers complained that they are requested to buy essential work equipment from X-platform’s app store, which is an unavoidable step in their sign-up registration process and prices are higher than market prices. As the initial start-up costs are unaffordable to many young rural migrants, they must accept the high interest rates charged on purchases, delivering X-platform workstations additional surplus from these workers. In addition, these purchases and debts further bind migrants to food delivery work. A driver described the situation:

An electric bike is around 3,000 yuan. The workstation manager will try to encourage new drivers to use loans to acquire a bike. They will charge 450 yuan each month for 12 months, which means you will pay an additional 1,800 yuan to get a bike. During this period, you can’t quit your job, otherwise you will pay breach of contract damages of around 3,000 yuan. For some young migrants who are new in the city, they have no choice but to accept these terms.

Recognising that drivers rely on food delivery work for their economic survival, X-platform has designed a salary structure that fosters and promotes the intensification of work. The salary structure has three parts: basic piece-rate income, additional bonus, and point-based rewards. In contrast to food delivery platforms commonly found in the West which use fixed rate piece-rate payments, X-platform employs an incremental piece-rate mechanism of step pricing which aims to motivate drivers and induce them to work as hard as they can. This system of pricing is depicted in Table 2. Under these workload metrics, the wage is calculated based on incremental payment tiers and the base rate of each delivery is always low. The pay of each delivery increases with the total quantities of completed deliveries. To seek a higher rate for each delivery, drivers must exceed the quotas by working around the clock. A driver remarked: “If you want a decent wage, you must work very hard every day without rest.”

Table 2. Drivers’ basic salary structure

Shift driver (full-time)
Quantity of deliveries per month	1–1,200	1,201–1,400	1,401–1,600	1,601+
Pay (RMB/per order)	4	5	6	7
Piece-rate driver (part-time)
Travel distance per	Within 3 km	3–5km	5–7km	Beyond 7km
Minimum bid price (RMB/per order)	4	4 + 1/km	6 + 1.5/km	13.5 + 2/km

Note: US$1 equals approximately RMB 7.

Source: Data compiled by the author based on fieldwork investigations.

The promises of bonuses is another mechanism which X-platform uses to intensify the work of drivers. Drivers are paid a bonus if they make a certain number of deliveries within a specified period. This includes “daily,” “weekly,” “monthly,” and “holiday” challenges. In the “daily challenge,” for example, if a driver completes 30 deliveries in a day, an extra 15 yuan bonus can be earned. A points game-based reward system, discussed in the following section, carries the same objectives to encourage drivers to meet higher delivery quotas. During fieldwork, I observed ways in which drivers strived to meet ever-higher quotas and some drivers even died from overwork.¹

X-platform’s exploitative salary structure contributes to intensified workloads. As Table 1 indicates, many drivers work for more than 12 hours a day, without a break, on weekends. This heavy workload, combined with strict time and quality control, may be a factor in high levels of traffic accidents among drivers. According to statistics, every 1.16 days a driver is injured or killed. Drivers made up about 60% of all traffic accidents reported in Shanghai in 2020.² As drivers are not employees of X-platform, the safety of drivers is not a consideration for the company. Rather, in seeking to ensure an even faster food-delivery service, X-platform has reduced the time limit for all deliveries to 30 minutes. A serious penalty, normally a whole day of income, is incurred for late delivery. To avoid penalties and achieve more deliveries, most drivers choose to take risks by driving at dangerous speeds so as to meet delivery deadlines. Almost all drivers stated that they have no choice but to race for a delivery by adopting dangerous driving behaviours. During fieldwork, it was common to see drivers violating traffic rules to make up time, such as driving against traffic and running red lights. Indeed, it was not uncommon to witness or hear of a tragedy that had happened to a driver, where they were injured or even died on the road. Of course they never received reasonable compensation. As one driver explained: “They [X-platform and workstation manager] act as our boss. We are working for them and are required to follow their instructions. However, once you get injured, they say we are not working for them, leaving us to take all responsibilities. You can’t get a cent from them.

Predatory Algorithmic Control

Based on these employment relations, X-platform uses additional algorithmic technologies to extract surpluses from drivers and control the labour process. Algorithms constitute a despotic form of control over drivers, and buttress asymmetrical relations of power that exist between drivers and the platform. Three factors were identified as key to X-platforms’ management of work: first, the algorithmic labour process; second, the gamification of the labour process; and third, digital-based performance evaluation. As outlined below, each of these generate further insecurity and uncertainty for drivers.

Relying on algorithmic technology, X-platform has designed two essential digitally enabled systems to instruct, co-ordinate, and monitor the whole food-delivery process: AI-powered dispatching system and AI-powered delivery assistant. These advanced systems achieve a comprehensive automation of the food-delivery process. Drivers are subjected to a virtually networked digital assembly line and are compelled to follow step-by-step algorithmic instructions which effectively strips them of control of the labour process. The entire food-delivery process is split into a series of tiny and repetitive tasks. Meanwhile, by means of information asymmetry, drivers are compelled to complete each task according to orders from algorithms under digital surveillance. Drivers claimed that their labour activities are under the absolute control of the platform. Drivers become cogs in the machine, unable to develop vocational experiences, skills, or increased competencies. They remain cheap disposable workers who can be easily replaced. The high degree of digitised automation of the labour process leads to further work deskilling and devaluation. Moreover, it sets low entry barriers to driver delivery work and undermines workers’ bargaining power. For instance, after the outbreak of Covid-19, many furloughed or laid-off workers from other industries surged into food-delivery work. This posed severe threats to the income of existing full-time drivers (see Huang 2021). As one driver remarked:

We are like robots driving through the streets … We are just puppets under the control of the platforms. We need to work very hard and be submissive. We can’t dare to argue for fair pay, that is almost impossible. Otherwise, they will deactivate or even cancel our working account … I really don’t know what my future would be if I lost my job … For the moment, although the working conditions are very bad, I still have a survival wage.

Digitalised performance evaluation is another tool leveraged by X-platform to exploit and dominate drivers. On the one hand, the platform deliberatively creates a vague performance management system to take benefits from drivers. Many drivers complained that their actual workload is usually subtly reduced by algorithms as the pay for each delivery is calculated on the distance driven. My own experience was that the algorithm’s frequently confirmed distance was below the actual distance driven. Moreover, the utilisation of customer review and rating systems, in favour of customers, constrains drivers to passively accept the prolonged uncompensated labour time spent actually making deliveries. Customer reviews are directly associated with rewards and penalties for drivers. A low rating or complaint from a customer can result in a penalty of 200–500 yuan, equal to one or two days of hard work for drivers. The actual amount of the fine is at the “discretion” of the programmed algorithms, which remain opaque to outsiders. Some customers abuse their rating power to request drivers to do extra work and sometimes make other unreasonable demands. This means drivers must do everything they can to satisfy customers’ demands to avoid capricious and malicious comments. A driver explained:

Customers often request us to help them buy additional products in other stores, such as cigarettes, a box of beer, bottled water and so on. Sometimes they ask me to help them to throw out their home rubbish. We are really in a dilemma, because if we refuse to do these things, the customer may give us a bad review or make a hostile complaint; if we accept, this will likely delay our next delivery …

X-platform also promises customers that the bill for the delivery of their order can be waived as compensation for late delivery. To take advantage of this, some customers have come up with various ways to hinder the driver’s completion of a delivery on time, such as changing their delivery address midway through delivery. In addition, to ensure the quality of service, X-platform initiated a “Smile Campaign” which demands drivers smile while working. This subjects drivers to the vagaries of algorithmic monitoring as, once they receive a random inspection alert, drivers must stop working and let their phone camera scan and validate their “smiling face.” In brief, digital performance evaluation shifts all of the risks and responsibilities to drivers.

In an effort to induce drivers to work harder, X-platform “gamifies” the food delivery process. One key outcome of the use of game-like features is that drivers often appear to consent to their own intensified self-exploitation, a pattern identified by Burawoy (1982). Game elements and principles underpin the food delivery process, encouraging young drivers to view work as “play.” There are several aspects to this. First, the delivery process is designed like a “pass through” or “check point” game. The progress of each task is tracked on a phone screen which is designed to resemble a game. Each delivery, from initial order request through to confirmation of delivery, is broken into a chain of discrete tasks. Drivers are guided step-by-step through each task via an algorithm. After completing each step, drivers report and then are notified of their next task. Task completion is tracked in real time. One driver reported to me that delivering food was like the “Temple Run” racing game, and he felt that he was being pushed to work harder and harder.³ Second, X-platform uses a points-based ranking system to motivate workers (see Table 3). Drivers are classified into different categories of “Knight,” an appellation that references a top-selling mobile game called “Honour of Kings.”⁴ The achievement of a higher “Knight” ranking means a driver receives more profitable orders, increased pay for each delivery, additional bonuses, and other privileges. Points are accrued for the number of monthly deliveries, good ratings from customers, and for working in adverse weather conditions. Points can also be deducted for low productivity, poor customer ratings, and late delivery. This system traps drivers in a seemingly endless effort to pursue and maintain ever-higher rankings. As one driver observed:

Table 3. Drivers’ ranking, points, and reward

Rankings	Points	Reward (RMB/per order)
Bronze knight	0–999	0.0
Silver knight	1,000–1,999	0.3
Gold knight	2,000–3,999	0.4
Platinum knight	4,000–7,999	0.5
Dimond knight	8,000–14,999	0.6
King’s honour	15,000 and above	0.7

I can say that I sacrifice my life to race for deliveries … If I want to be paid better, I must achieve the ranking of Diamond Knight or above. This means that I must complete at least 40 deliveries per day on average and not make any mistakes. At first, I felt motivated by the growth of points and Knight level, however, after several months of delivering, it is really became a painful struggle.

Subservient Citizenship

Apart from the work precarity produced by platforms, the national and local states’ policies towards rural migrants makes the work and lives of drivers more precarious. The various types of work precarity drivers confront are reinforced and amplified by the second-class citizenship of the rural migrant. Drivers are thus in a discriminatory social labour environment that generates additional insecurity and uncertainty.

The well-known hukou system means that many drivers from rural areas are not entitled to full citizenship rights in urban areas.⁵ Drivers with rural hukou struggle with unequal access to public service and welfare in terms of housing, the public medical service, and education. There are several ways in which this occurs. First, the system denies migrant workers access to affordable and sanitary housing. For example, with a plan to build a “socialist” metropolis, the Shanghai authorities re-developed those inexpensive districts in the city where rural migrants lived, forcing them to move into more expensive apartments. This meant that drivers have to pay an ever-greater proportion of their income on rent. As drivers are the main breadwinners for their families, they are forced to scrimp and save so they can send remittances to their rural-based households. During fieldwork, I lived with seven drivers in a one-bedroom flat, where seven bunks were crammed into the single bedroom and entrance hallway. Local authorities were not interested in these circumstances and made attempts to help drivers with their accommodation. Indeed, they considered such crowded accommodation to be “illegal.” Inspections by urban management officers meant that drivers often experienced great uncertainty in terms of their living circumstance and were often forced to move their place of residence.

Medical costs are a second factor threatening the income security of drivers, as drivers are not entitled to public medical insurance and hospital services in cities. Instead, drivers must pay much higher fees than urban citizens in accessing medical care. The problem is that under the compound effect of the nature of food delivery work, the salary structure, and the tighter time limitation of algorithmic control, the potential to be involved in a traffic accident is high. One driver said that “a normal traffic crash can cause me to go bankrupt, but it looks like it is avoidable.” Moreover, most drivers have to endure separation from their children who are studying in the countryside. At the policy level, the Shanghai government has recently permitted the children of migrant workers to access public education in the city from elementary to high school. However, in practice, most drivers say that their school applications are declined with “some ridiculous excuses” due to their identity as rural migrants. One driver shared his experience of discrimination, stating:

I attempted to send my son to elementary schools near where I work. At all the schools I approached, I was asked to fill out various forms in relation to my Hukou status as well as my occupation and provide relevant documentation. I think that because of my [rural] Hukou status and my job as a food delivery driver, none of the schools accepted my application claiming that all student places were full. I am not alone; all my fellow drivers suffer like me.

Drivers are alienated from, and discriminated against by, urban citizens. They are viewed as second-class people and labelled as a “low-end population.” Drivers are subject to a humiliating social and working environment and this also contributes to the precarious nature of their lives. The social bigotry against drivers carries with it various indignities, including workplace bullying. In some building or public areas, gates are emblazoned with notices such as “no access for drivers” and “drivers stay out.” One driver explained how social discrimination made the work even more difficult: “Every time when we [drivers] pick up order from Starbucks, the staff don’t allow us to wait inside the shop and we have to wait outside. The problem is that, although orders may be ready to collect, they [the staff] forget or intentionally fail to inform us. This delays delivery.” Another driver reflected on a humiliating experience he had: “A customer ordered me to bend to my knee in front of him. He threatened me with a negative review if I didn’t. This is because I rang the doorbell and woke his sleeping baby. I had no choice but to do as he said. I didn’t want to face the penalty incurred by his complaint.”

Apart from these forms of bullying and intimidation, physical attacks on drivers are not uncommon. Drivers say that they face risks of attacks by customers, (security) guards, and even traffic police. Figure 1 shows a young food delivery driver, who was knocked down by a woman in her luxury car and was then punched and kicked. As a driver said to me poignantly, “We are migrant drivers with little power and are ‘disposable’ … that is why they can bully us. Who can I blame? I can only blame the fact that I was born a rural peasant. It is like an original sin.” Another statement by a driver reflects generally held feelings of grievance:

Figure 1. A driver being attacked

Note: The picture was taken by the author during the fieldwork; personally identifiable information is pixelated.

The platform only cares about whether the food is delivered according to their standards. The customer only cares about having their food delivered on time. The traffic police, who are supposed to protect us, only care about whether they can issue us with a fine … Nobody cares about us. They all want to extract some benefit from us … They have nothing to fear because they know we can’t survive without delivering food in the city … To be honest I think the delivery of food is seen as being more valuable than my life.

Suppressed Labour Resistance

The lack of proper employment regulation, algorithmic control, and second-class citizenship combine to produce the conditions for precarious work experienced by delivery drivers. Drivers confront significant obstacles in trying to resist their exploitation, at both individual and collective levels. As outlined below, both the state and platforms play critical roles in constraining driver resistance.

At an individual level, drivers have developed various tactics to resist controls imposed by the platform. However, many of these tactics cannot be sustained, and have been successfully countered by the platforms as well as political authorities. For example, to achieve greater benefits as well as more autonomy at work, one tactic used by drivers has been to install “plug-ins” (software components) on their phones to by-pass the power of the algorithm. One programme or “bot” that can be used alters the GPS location. This helps drivers avoid the platform track-and-trace system. Drivers can then make their deliveries via shortcut routes that they know about. Another bot used will automatically filter orders which enable drivers to select those orders for which the pay is higher. Research has suggested that the use of bots can be a critical means through which drivers can exercise individual resistance (Chen and Sun 2020, 1574). However, fieldwork data suggest that this tactic is strictly limited by the platform. First, the platform uses its technological superiority to constantly update its applications and algorithms to prevent drivers from using bots. Second, platforms also punish drivers caught using bots. If discovered, the driver’s working account with the platform will be permanently cancelled. In extreme cases, the platform will take the driver to court where they will be accused of disruption of the market order and fair competition.

Another tactic used by drivers has been to try to use the legal system to counter unfair treatment by the platform. For example, some drivers have taken the platform to court and local government authorities claiming that platforms should treat drivers as formal employees and thus they should have access to worked-related benefits and welfare. However, despite continued attempts, none of these endeavours have been supported by the authorities. Indeed, despite being denied formal employment status, reasonable appeals by drivers are usually rejected by local authorities. One case encountered during fieldwork concerned a driver who lost a leg in a traffic accident while working. The platform denied any responsibility, giving the driver a mere 2,000 yuan as “humanitarian compensation.” The driver took his case further to various labour-related departments, but with no success. As one driver observed: “I do think it is a conspiracy. You know the possibility of drivers being involved in crashes is very high. But, even if you become disabled or even die, the court will rule in favour of the platforms and refuse to provide fair compensation…, claiming that we are not their employees.”

Drivers have also resorted to video activism to press their claims. Many drivers in Shanghai used Tik Tok to express their anger and air their grievances in the hope of damaging the reputation of the platforms. Drivers create short videos that show the precarious nature of their work, aiming to expose how algorithmic management is used by platforms to exploit them. The videos contain content that highlights risky working conditions, the intensity of workloads, illegal imposition of penalties and fines, the hostility of customers, and examples of others forms of unfair treatment. Some of those produced have become Tik Tok’s top trending videos and help trigger wider public criticism of the platforms. In response, the platforms have employed the “Internet water army” which posts comments that attack drivers that produce such videos as well as rejecting the claims made by drivers. The authorities also actively repress driver voices by using the internet surveillance and censorship system. Drivers have reported that their video posts are removed by the system, and their accounts blocked for alleged violations of the Internet Law.

Collective forms of resistance, in particular, are subject to severe constraints by political authorities. Online interactions between drivers and attempts to collectively mobilise are closely monitored. Because of the dispersed and individualised nature of food delivery work, drivers face significant difficulties in building collective agency. Recent work has suggested that social networking platforms such as WeChat can be used as tools for mobilising strikes and protests (Liu and Friedman 2021). Shanghai drivers have also used Tik Tok, WeChat and QQ in attempts to build virtual forms of solidarity. However, unlike these studies which argue that WeChat enables attempts to build collective resistance to fly “under the radar,” this research revealed that drivers’ online interactions and activities are under strict surveillance. Drivers claim that the uses of social networking applications are limited to developing connections, exchanging information, and helping to build mutual support. The authorities view labour movements as especially sensitive and, as a result, drivers say that they don’t dare to become involved in activities like planning strikes among their virtual groups. To do so runs the risk of having their group interactions immediately blocked, and group owners being investigated by the police.

One example of how attempts by drivers to engage in collective action are constrained by authorities is that of a strike for better pay and conditions organised by some drivers in Shanghai in 2019. The drivers had formed a digital group to plan the strike but organisers had only “hinted” at their tactics and aims to the group. Drivers’ participation in the strike was mobilised largely offline, occurring at driver work stations and other places where drivers could physically meet together. On the day of the strike, drivers assembled at six separate places in the city. Fearing repercussions, drivers did not shout slogans or display banners, choosing to gather in front of shopping malls and food courts where they refused to pick up orders. However, the collective action could not be sustained. The police intervened and ordered that the crowds disperse. The strike was quickly broken, and the strike leaders were arrested. Drivers reported that some of the virtual groups that had been formed were blocked. The main strike organiser was charged with “picking quarrels and provoking troubles,” a charge that carries up to five years’ imprisonment. One driver remarked: “The government always knows what you are thinking and what you are doing. I think our conversations on the phone are monitored and censored. If you want to plan a strike in any form, the police will find you and you may go to jail …”.

The Shanghai strike illustrates the inherent difficulties faced by drivers who “walk on a razor’s edge” in their endeavours to organise via online internet applications, and the political repression these attempts attract. Cheng Guojiang, a leading driver activist created several virtual groups on WeChat and QQ that connected more than 15,000 food delivery drivers. He was arrested in 2021 and all the groups that he created were blocked.

Conclusions

This article has investigated the ways that the precarious work of food-delivery drivers in Shanghai is produced and reproduced. The focus has been on the combined role of the state and the platforms’ management of the labour process. Three key determinants have been identified. First, the platforms are able to avoid their statutory and other responsibilities, reduce their costs, and avoid the business risks associated with sponsoring food-delivery work by outsourcing their labour and hiring practices to third-party labour recruiting agencies. Second, the platforms use digitisation, algorithmic technologies, and gamed-based applications to control and monitor the labour process in ways that intensify processes of surplus extraction. Third, the denial of full citizenship rights to rural migrants creates a pool of extremely vulnerable workers who have little choice but to accept low-paid, risky, and unprotected food-delivery work. Under these circumstances, the capacity of drivers to build and sustain individual and/or collective forms of resistance is limited and, when it does occur, is subject to repression by the platform company and central political authority.

This article makes the following theoretical and empirical contributions. First, by highlighting the Chinese state, the platforms, and their combined role in producing precarity, this article validates Alberti and colleagues’ (2018) argument that the state and management are two key drivers of precarity. In the Chinese food-delivery economy, the platforms take advantage of current regulatory loopholes of platform work, as well as the digital reconfiguration of the labour process, to create production-based precarity in the platform-based workplace for drivers. These types of precarity are further reinforced and amplified by the pivotal role of the Chinese authorities. This manifests in the interplay of institutional discrimination against migrant workers and the Chinese government’s indulgent policy towards the industrial digitalisation that shapes citizenship precarity for rural migrant drivers, who are forced into a hostile and discriminatory social environment that makes their work and life even worse in urban areas. These findings echo an assertation by Kalleberg and Vallas (2017) that digitalisation is a vital force for work precarity, as it allows the emergence of unbridled employment forms and new systems of management that institutionally favour the platforms.

Second, this article contributes to existing debates about work precarity in the platform labour regime by offering the experiences of drivers who make a living through platform work. In contrast to rhetorical framings of platform work that can provide autonomy and flexibility for gig workers, this article shows that Chinese food-delivery drivers experience severe domination and pervasive surveillance both in the workplace and at the wider society level. As the migrant drivers depend on food-delivery work to survive, they passively accept ultra-precarious working conditions in exchange for work opportunities. This resonates with the contention that “the re-arrangement of capital through online digital technology can reproduce new forms of dependency, surveillance and subjugation” (Alberti et al. 2018, 452). In platform labour regimes, the findings show that drivers’ lived experience of precarity is not limited to financial instability, work insecurity and inferior class identity. More critically, due to the asymmetrical power relations between platform capital and platform labour noted by Huang (2022), drivers have to tolerate the consequences of the loss of the power of control in the labour process, shown in the ways their resistance is constrained both at individual and collective levels. This finding reinforces the claim that platform precarity should be understood in capital–labour relations that are embedded in the platform-based workplace (Vallas and Schor 2020).

More recently, and in contrast to previous policies that aimed to foster the growth of the digital economy, Chinese authorities have introduced new regulations to prevent what they call “monopolies and [the] disorderly expansion of capital” (XinhuaNet, March 5, 2021). On July 26, 2021, the Chinese State Administration for Market Regulation, along with six other administrative departments, issued a new national policy and guidelines to protect food delivery drivers (XinhuaNet, July 26, 2021). This new policy requests platforms to ensure food-delivery workers earn no less than the local minimum wage and participate in social insurance. It also requires that the “strictest algorithm” should not be used as an assessment requirement and unreasonably tight deadlines for deliveries should be appropriately relaxed. All delivery workers should be able to join labour unions. These changes suggest there may be some possibility for improving the working conditions of food-delivery drivers as part of China’s adoption of its broader “common prosperity agenda.” However, by late 2022, the authorities have yet to specify how the new rules will be enforced or what the penalties will be for platforms that violate the guidelines. More importantly, some of the fundamental causes of precarious work, especially outsourcing and restricted citizenship rights, have not been addressed at the policy level. It seems that China’s drivers will continue to be denied access to basic labour rights, social protections, and will also continue to face persistent precarity and other hardships associated with their work in the food-delivery industry.

Acknowledgements

I am very grateful for the constructive, professional, and insightful comments from the journal’s editor and anonymous referees, one of whom provided remarkably detailed and helpful comments. I also thank Ye Liu for her valuable comments on earlier drafts of this article. Finally, I owe my food delivery fellows in Shanghai a great debt of gratitude for their help.

Disclosure Statement

No potential conflict of interest was reported by the author.

Additional information

Funding

This work was funded by China Scholarship Council (Grant No. 201808310053). The data collection and analysis were supported by UCCL Grants from Universities’ China Committee in London, a Fieldwork Grant from Gilchrist Educational Trust, Henry Lester Trust, The Great Britain-China Educational Trust (GBCET), and a Travel Grant from Department of International Development at King’s College London.

Notes

1 There is extensive media coverage of Chinese food-delivery drivers who are said to have died of overwork (see, for example, Daily Mail, January 18, 2021).

2 The statistic is based on a report from official WeChat account of the Shanghai Municipal People’s Government Information Office. Details can be found in: https://mp.weixin.qq.com/s/4TcJdPV9YywfHCoegxySxA.

3 Temple Run is a popular mobile game in China. The player controls an explorer who must keep running without any breaks to avoid a demonic monkey that is in chase. The explorer collects golden coins in the process of running and the number of golden coins depends on the running speed and distance.

4 Honour of Kings is the highest-grossing mobile game worldwide. There are eight large matchmaking tiers, namely Tough Bronze, Order of Silver, Glory of Gold, Noble Platinum, Eternal Diamond, Master Star, Super King, and Honour of Kings.

5 The Hukou is the household registration system that determines access to various social benefits and social protections. It also results in an economic-political differentiation between holders of urban and rural hukou, with the latter institutionally disadvantaged when they live sand work in an urban area.