Beyond Precarity: Forced Labor in China’s Ride-Hailing Industry

Abstract

Scholarly discussions of the consequences of the on-demand economy on work mainly focus on precarity. Using a case study of Didi Chuxing, this article moves beyond this conventional approach to highlight coercion as a striking feature of labor relations in China’s ride-hailing industry. Drawing upon the conceptual tool of neo-bondage, this article foregrounds the central role played by forced labor in securing a cheap and docile work force during Didi’s rapid market expansion. This article advances the existing literature in two ways. First, it highlights the need for a more robust analysis of the productive forces in the on-demand economy. Second, it argues that the on-demand economy not only represents an intensification of the ongoing trend toward precarity, but also an extension of forced labor regimes from electronic assembly lines to the service industry.

Keywords

Didi Chuxing forced labor instruments of labor neo-bondage on-demand economy ride-hailing industry

Introduction

The growing on-demand economy continues to have its critics and its skeptics. At the core of these debates are employment practices. A growing literature provides empirically grounded and insightful accounts of on-demand labor. Key topics include how work is redefined and reorganized (Hyman, 2018; Prassl, 2018), how workers are managed (Griesbach et al., 2019; Rosenblat, 2018), and the various ways in which laborers respond to the managerial process (Chan, 2019; Chen, 2018). What seems to bind together most of the critics is the tension between the classification of workers as independent contractors and the platforms’ unprecedented control over the organization of work. On one hand, platforms designate service workers as independent contractors without a formal contract (Stanford, 2017). On the other hand, platforms leverage multifaceted managerial techniques in the labor process, including calculative authority (Shapiro, 2020b), algorithms (Rosenblat, 2018), communicative technologies (Li, 2021), organizational and legal control (Lei, 2021), and so on. Taken together, the existing literature is replete with arguments regarding the consequences of the on-demand economy for labor, either from a macroeconomic perspective or looking at the long-standing questions of labor control and resistance.

This work is highly illuminating, yet I highlight an important gap in our knowledge about the production of on-demand services. Although material or matter—what I mean here are labor instruments in the Marxian sense—is clearly the infrastructural basis for the function of the on-demand economy, it remains underexplored in the existing literature. For example, ride-hailing drivers must have a vehicle before driving; food-delivery couriers need means of transportation—vehicle, motorcycle, or bicycle—that grant them the right to deliver. Explanations for this negligence are multiple and complex. However, the notable one is an implicit assumption underlying the existing corporate discourse that on-demand platforms make full use of individuals’ assets which are otherwise spare, a key rationale behind the so-called sharing economy (Schor, 2021; Shapiro, 2020a: 97).

Nevertheless, there are points of conflict between the real sources of labor instruments and platforms’ accounts of them. As the article highlights, platform companies not only commodify individuals’ spare possessions but also soak up a mass of cheap laborers at the precarious edge by providing them with labor instruments on credit. For example, food-delivery platforms expand delivery work to disabled people by providing them with mobility equipment; ride-hailing platforms offer car leases to people who otherwise would not be eligible to drive. The reality of platforms as the providers of labor instruments requires us to revisit not only the corporate discourse of the sharing economy but also the labor relations between platforms and service workers.

This article departs from the existing literature by giving the instruments of labor due account. Focusing on the vehicle-leasing programs of Didi Chuxing, China’s ride-hailing giant, this article explains debt-based forms of organization and production and their consequences for labor relations. Inspired by Jan Breman’s work (Breman, 2010, 2014), this article defines the modality of employment to which drivers are recruited as “neo-bondage”—a term for series of manifestations of bonded labor in the contemporary informal economy. Drawing upon interdisciplinary scholarship from sociology, anthropology, and media studies, this article describes how the neo-bondage labor regime plays out in China’s ride-hailing economy, the impetus for it, and an overall hierarchical set of power relationships and interdependences it engenders. I argue that debt serves as a mechanism of labor control that enables Didi to secure a “dependable and disposable” (Lei, 2021: 281) labor force and externalize regulatory risks, which nevertheless brings about forced labor practices that severely threaten drivers’ wellbeing.

Bringing forced labor to the fore, this article complicates the existing literature on the gig economy and platform labor, which tends to focus on precarity. Discussions on labor issues in the gig economy emphasize how temporary work accelerates the already existing insecurity and vulnerability for a significant population of laborers (Hyman, 2018). However, beyond precarity, coercion is another troublesome feature of the contemporary informal economy (Hatton, 2019, 2020), which has not yet received enough attention. As demonstrated in this article, when Didi drivers are trapped in debt, what bothers drivers is not only whether they will lose the means of livelihood—a question afflicting all laborers in precarious jobs—but also the fact that they are subject to various forms of coercion that threaten individual freedom and wellbeing. By emphasizing coercion, I do not deny the importance of precarity. Quite the opposite, following Hatton, I emphasize that precarity and coercion are connected in labor relations because flexible employment requires docile labor forces and “coercive labor regimes prepare productive and obedient labor forces for future exploitation in the low-wage precarious economy” (Hatton, 2019: 2).

Literature review

Defining and debating the uberized economy

Beginning from around 2008, a wave of new businesses emerged to provide efficient on-demand services through connecting service requesters (consumers) and providers (workers). These businesses are based upon the application of algorithms that enable a good match between the two sides and also exert control over workers’ service qualities (Kenney and Zysman, 2016). At the leading edge of this wave is the ride-hailing company Uber. Unlike traditional cab companies that own vehicles or employ drivers, Uber acts as the so-called digital intermediary that offers instantaneous connections between passengers and drivers. The extant literature provides fruitful discussions on the role of algorithms in Uber’s daily business (Rosenblat, 2018). For example, Rosenblat documents well how Uber leverages algorithmic control—characterized by informational asymmetries, gamification, and being fickle and opaque—to maneuver drivers (Rosenblat, 2018). The success of Uber inspired various companies across industries—ranging from ride-hailing, food delivery, medical care, and education, etc.—to simulate its business model, and this commercial trend has been dubbed the Uberized economy (Schor, 2021).

A variety of concepts in academia have emerged to encapsulate the essence of the Uberized economy. Business scholars and technological enthusiasts offer the concept of the sharing economy (Sundararajan, 2017). Based upon peer-to-peer informal exchanges, the Uberized economy promises an efficient utilization of idle resources and a new way of working characterized by flexibility and autonomy (Rosenblat, 2018). Nevertheless, critics from sociology, law, and media studies provide compelling challenges to the rhetoric of sharing. Central to their criticisms are two issues. First, Uberized companies leverage the positive meaning of sharing to gloss over their labor practices (Schor, 2021). Scholars of labor studies replace the “sharing economy with the ‘gig economy” to highlight the piecework offered by Uberized platforms (Ravenelle, 2019). Second, the language of sharing enables Uberized companies to claim their technological exceptionalism, which in turn helps them evade regulatory frameworks. Ride-hailing platforms, for example, provide the functional equivalent of the taxi. However, by characterizing themselves as a sharing company that merely offers matching algorithms, they avoid many of the regulatory requirements for taxis (Prassl, 2018; Rosenblat, 2018).

Beyond this conventional critique, this article emphasizes that the language of sharing in many ways encourages a de facto neglect of more obviously material phenomena and processes; that is, the neglect of thingness in labor relations. Attending to the question of labor instruments, a few scholars argue that the Uberized economy should be understood as “a digital reincarnation of the putting-out system” (Acquier, 2018), a preindustrial organizational form that even precedes the emergence of full-fledged capitalism (Kenney and Zysman, 2016; Stanford, 2017). Putting-out is a system of manufacturing in which merchants supply raw materials to artisans (or so-called independent craftsmen) who use their own instruments of production to convert them into finished goods at piece work prices. Artisans need to return the finished products to merchants who are responsible for sale (Baishya, 1997). At first sight, the putting-out economy is an appropriate description of the Uberized economy. For example, ride-hailing platforms take command of information that dispatches orders to drivers, arranges routes, and determines fares; drivers transform the information provided by platforms into transportation services (Rosenblat, 2018).

However, I emphasize that the “putting out” analogy is still insufficient to account for the phenomenon, because it neglects a trend wherein platforms provide options to loan instruments to workers. As discussed later, when drivers rent vehicles to work, they don’t use their own tools to produce ride-hailing services. Instead, they are bonded to some sources of debt, and they work off the debt. Thus, the existing discussions—either the camp of the sharing economy or the thread of the putting-out economy—fail to capture a group of workers who had no materials to begin their work and had to buy on credit or rent from the people they worked for. Capital is a social relation between people established by the instrumentality of things (Marx, 2014). Therefore, to understand labor relations in the Uberized economy, it becomes pressing to incorporate the relative importance of the productive forces into the analysis.

Theorizing debt bondage and forced labor

While slavery was abolished in most countries during the 19th century, scholars from sociology, history, and anthropology emphasize debt bondage as a key feature of present capitalism (Guérin et al., 2015: 12). In debt bondage, individuals borrow money or assets and pledge their labor against the loan until debts are repaid (Kara, 2008: 221–222). However, due to the manipulation of debt, credit, or contract by landlords, employers, or intermediaries, bonded laborers find it very difficult to struggle out of the relationship and have to continue to work involuntarily (LeBaron, 2014: 766). The economic model of debt bondage dates back to centuries in various forms, such as peonage, serfdom, and sharecropping. In the traditional forms of debt bondage, obligations are often hereditary, and family members of debtors can also be trapped in the relationship that serves masters (Breman, 2010; LeBaron, 2014).

Moving beyond this age-old form of debt bondage, anthropologist Jan Breman proposed the model of “neo-bondage” to better explain the bonded relationship in the informal economy in recent decades (Breman, 2010, 2014). On one hand, neo-bondage resembles historical practices of bonded labor. In general, both forms start from indebtedness, and indebtedness serves as a mechanism of dependency. Debtors must repay the debt in labor until the debt is paid off. On the other hand, neo-bondage differs from the older forms of bonded labor in that the relationship between employers and workers is less personalized, more contractual, and of a shorter duration, which often links to a cycle of production (Breman, 2010). In neo-bondage, the oppressor is often a capitalist entrepreneur that exploits labor to pursue profit and accumulation (Guérin et al., 2015). Therefore, as opposed to the more complex social and personal dependency manifested in the traditional forms of bondage (Kumari, 2018), neo-bondage is primarily based upon economic transactions where debt leads to involuntary service.

All debt bondage invariably contains some elements of forced labor because indebtedness serves as a mechanism of dependence that significantly weakens workers’ freedom of choice. A rich body of literature has examined the phenomenon of labor coercion in debt bondage across low-wage labor sectors (Breman, 2010; Guérin et al., 2015). As research demonstrates, at the core of forced labor is its involuntary nature, which means workers are held against their will and coerced to comply under the threat of penalty (Shields, 2011: 177). But involuntariness and coercion can emerge at various points in the labor process (Strauss and McGrath, 2017). Of particular scholarly interest is the role of deception and fraud in the recruiting stage (Skrivankova, 2010). That said, workers are not necessarily forced to enter a labor relationship under duress. Many often take up work voluntarily; however, they gradually discover that they have been deceived about the conditions and nature of the work but are no longer free to leave without repercussions (Skrivankova, 2010). Prior research also highlights the fact that the forms of coercion used by employers are multiple and not easily recognizable (McGrath, 2013). Physical coercion—such as visible signs of physical restraint—might not often be used by employers. More common forms include economic coercion—for example, by not being paid the promised wages and threatened with loss of work—and social coercion, which refers to various legal but highly punitive sanctions that threaten workers’ wellbeing and future (Calvão, 2016; Hatton, 2019). To sum up, all this research broadens the scholarly understanding of the features of forced labor and its connections to contemporary capitalism. Building upon the above insights, this article will analyze neo-bondage in Didi’s vehicle-leasing programs.

Case selection, data, and methods

Throughout this article, I use Didi as shorthand for DiDi Chuxing, for narrative ease. Didi is China’s ride-hailing giant. Since its founding in 2012, the company has been marked by its spectacular growth. In 2016, 4 years after its inception, Didi became the world’s largest ride-hailing company, with 25 million trips a day in China, surpassing the combined daily trips of all other ride-sharing companies across the globe (Zhu and Iansiti, 2019). Up until very recently, Didi held an estimated 90% share of the domestic ride-hailing market. In 2021, the company raised US$4.4 billion in its US initial public offering (IPO), making it the biggest Chinese IPO in the United States since 2014 (Chen, 2021). Didi offers multiple services beyond ride-hailing, such as bike-sharing, food delivery, finance, and so on. However, ride-hailing constitutes the company’s core business (Qi and Li, 2020), making it the focus of my research inquiry. Most of the research on Didi by media scholars has two strands: one focuses on drivers’ working conditions and experiences (Chen, 2018; Qi and Li, 2020), and one focuses on the company’s data strategies and its relationship with the state (Chan and Kwok, 2021; Chen and Qiu, 2019). This work is important, but it barely approaches the subject from the perspective of production and lacks firsthand information from the corporate side. As Didi is situated at the center of designing and implementing labor strategies, such limitations hinder a thorough investigation of the implications for drivers who work for the company.

This study adopts a qualitative research methodology that entails a hybrid of semi-structured in-depth interviews and observations in Beijing. Over the course of my fieldwork in 2018 and 2019, I conducted 68 in-depth interviews, including with 15 employees of Didi, 50 drivers, and 3 managers of fleet management companies (hereafter, FMCs). First, my interviewees at Didi are characterized by their diversified backgrounds, including rank-and-file big data and algorithmic technicians, senior technology officers, product managers, and employees from the departments of policy studies, academic collaborations, marketing, and public relations. Conducting interviews with people from China’s leading Internet companies is not easy, but it is possible. I drew on personal connections to access interviewees while making clear the voluntary nature of the study to ensure there is no sense of obligation, due to any potential existing relationship we may have. I also visited Didi’s headquarters in Beijing twice, and both times, I was invited by its senior managers. The two visits enabled me to observe Didi’s working space, which gave me an impression of the corporate ambience and culture.

Second, I made observations with over 50 drivers by riding as a passenger in Beijing. They were all drivers of Didi Express, the most economical and popular mode of Didi services. I conducted interviews with them either in their cars or by phone. I assured drivers that whether they were willing to accept my interview would not affect my rating of their services. All these interviews were voluntary and anonymous. Among all the drivers I met, only two were part-timers: one was a retired professional who had a regular pension, and the other was a stay-at-home mum. All the others that I met were full-timers: one was a Beijing local who came from a remote rural district, and all the rest were non-local drivers. Most drivers were aged between 20 and 50 years old, and only one was a female. While qualitative research rarely desires samples that would be necessary for achieving generalization (Tracy, 2019: 268), the gender distribution revealed in my sample can still powerfully resonate with multiple contexts. In current studies about China’s ride-hailing industry, there lacks an authoritative estimation on the ratio of male drivers to female drivers at the national level. However, scholars, drawing upon survey data in different locations, have consistently highlighted the extremely disproportionate gender ratio in the industry, with less than 5% of female drivers (Chen, 2018: 2695; Qi and Li, 2020: 510).

Third, I conducted interviews with managers of three large FMCs in Beijing. These interviews were conducted remotely and anonymously to protect the safety and privacy of the interviewees. Access to FMCs was enabled by my long-standing connections with Chinese journalists who have done reports on the ride-hailing industry.

This article adopts an interpretive paradigm with the belief that “both reality and knowledge are constructed and reproduced through communication, interaction, and practice” (Tracy, 2019: 51). Thus, instead of arguing that there is an absolute impartial reality out there, I emphasize reality is always mediated through people. To ensure the validity of this research, I draw upon “methodological triangulation,” that is, I use multiple sources of data to support my arguments (Tracy, 2019: 50). As this article is only one part of a larger project on China’s platform economy, it does not present all the interviews that I conducted, but my analysis here is informed by them. I analyzed the interview data based on the phronetic iterative approach, alternating between firsthand data and conceptual frameworks (Tracy, 2019: 6). Multiple themes and codes emerge from my interview data, among which coercion is a salient one among drivers who rent vehicles to drive. As this article pays attention to the materiality of on-demand labor, my focus is given to the relationship between vehicle-leasing programs and the labor practices involved. In presenting interview data, I have adopted the combination of using the informant’s own words and my paraphrasing. For some important quotes, I prefer to use the informants’ own words, which can give the texts directness and vividness; for some recurring answers, I combine and paraphrase them to ensure the clarity and flow of the texts. As an effort to provide both professional and social anonymity, I use pseudonyms in the article.

Findings

Vehicle-leasing programs: prototyping neo-bondage relations

Didi began collaborating in 2016 with fleet management companies (FMCs) across China that lease vehicles to prospective drivers. The specific terms of different programs vary from place to place, but they boil down to two dominant formats: drivers either sign contracts with FMCs to finance a car or to lease a car. Both require drivers to provide ride-hailing services for a fixed period, ranging from 6 months to 3 years.

Launched on 18 April 2016, the “partner entrepreneurial project” (huoban chuangye jihua in Chinese) serves as the prototype of the recurring neo-bondage relationship. The project was designed to recruit 10,000 full-time drivers in five cities: Beijing, Guangzhou, Shenzhen, Wuhan, and Chengdu. As Didi marketized, drivers could gain a new vehicle chosen from three designated models as long as they paid an amount of money for an upfront deposit. Depending on different brands, the deposit ranged from 15,000 RMB to 20,000 RMB. After providing ride-hailing services for Didi for 2–3 years, drivers can either request the deposit be refunded or keep the vehicle for personal use (Tencent Technology, April 18, 2016). To promote the project, Didi created eye-catching slogans such as “drive to own” or “get a new vehicle for free.” Didi also promoted the project as socially good because it offered great opportunities for the unemployed or underemployed to “earn a monthly income above the city average in the five cities” (Tencent Technology, April 18, 2016).

At first glance, the partner–entrepreneurial project was genuinely attractive: quick and accessible work that earned a decent income and simultaneously provided a vehicle for personal use. In addition, Didi’s language of “partners” and “entrepreneurs” conveyed a feeling of mutual help and generosity that would provide drivers with a true chance to prosper. In fact, some drivers mentioned that they were convinced by Didi’s advertisements to join, with the hope of having better economic opportunities.

However, what is startling about the campaign is that it involves deception, enticing individuals into a cycle of indebtedness and involuntary servitude. In the recruitment process, Didi insisted that no additional monthly payment would be required during the contract, which nevertheless turned out to be false in terms of benefits (Tencent Technology, April 18, 2016). In fact, the essence of the project was a predatory mortgage whereby Didi loaned vehicles to drivers with an exorbitant rate of interest. In addition to the one-time deposit, drivers must undertake a heavy workload and pledge a percentage of their earnings to Didi. Yet, the darker side of the project has been consistently hidden from public view. Chen, a 26-year-old driver, explained the revenue-sharing scheme imposed on him:

I am required to achieve a minimum monthly gross income of 8000 RMB, an amount after the deduction of an approximate 20% platform fee. If my gross income is equal to or less than the threshold, I need to pay 8000 RMB as rental. If more than 8,0000 RMB, Didi will charge me 6000 RMB as rental, but I also need to pay it 10% of the part exceeding 8000 RMB.

The vehicle Chen rented was worth about 70,000 RMB. Simple calculation can quickly reveal that the total rental accumulated in 3 years is about twice the market price of the vehicle. However, not only are the loans predatory in the economic sense—far more than typical financing—the debt burden is transformed into a form of economic coercion which Didi uses to compel drivers’ compliance. In Chen’s retelling of the scheme, he repeatedly emphasized the importance of fulfilling the output requirement, because the consequences of not doing so are severe. Chen said, “I must make up the difference by using my own money.” Chen’s scenario was common among drivers enrolled in the partner–entrepreneurial project.

The debt burden subjects drivers not only to a state of economic dependency but also involuntary servitude, whereby drivers’ choices and bargaining power are severely restricted. Debt becomes a mechanism of labor governance through which Didi cultivates a dependable and disposable workforce. Didi developed unique requirements on enrolled drivers that distinguished them from others driving with their own cars. Unique is a rather vague term, but it is difficult to be precise given the diversity and constant evolution of rules imposed on enrolled drivers in different times and places. To understand the conditions of forced labor, I emphasize three indicators. The first concerns working hours. Didi imposed the required workload on enrolled drivers, resulting in them working full time, or perhaps more precisely, overtime. Thirty-one-year-old Lin said, “I am required to hit a minimum of 24 rides per day and drive at least 6 days per week.” When I asked how long he drove to meet the requirement, he said, “it depends; I had good days and bad days. But in general, I drive 8–10 hours per day.” Twenty-seven-year-old Zhao revealed, “I am required to drive with passengers for at least 6 hours a day; if adding empty-cruising hours in, I need to drive 8–9 hours to meet the requirement.” Some drivers mentioned that they felt they were tricked by Didi. Thirty-five-year-old Song said, “That is a cruel hoax. Driving at such intensity for three years, the so-called drive-to-own is nothing more than an empty check because vehicles must be decommissioned because of high mileage.”

Although enrolled drivers have to work full time, and most of the time, overtime, these drivers are not deemed to be Didi’s employees because they were recruited on a contract system through FMCs. In some but not all situations, enrolled drivers were required to sign a labor service agreement (laowu hetong) with their FMCs the moment they gained the new vehicle. Unlike a labor contract (laodong hetong), the labor service agreement in China is applied to a situation where the relationship between employers and workers is mediated by a range of labor intermediaries or contractor brokers. The contract is short-term and informal, and workers are classified as independent contractors without being entitled to any insurance and benefits. However, work under this extremely flexible contract tends to be highly exploitative because of the absence or ineffectiveness of legal enforcement and state supervision (Friedman and Lee, 2010: 512).

The second indicator involves the restraint imposed on drivers’ choices. The standard logic of ride-hailing work is that drivers can determine when and where they drive. While platforms will nudge drivers’ behaviors through incentives that subtly undermine their flexibility, drivers still, in theory, have the room for choice (Li, 2021). However, Didi imposed compulsory requirements on drivers’ temporal and geographical locations, which exerted tremendous economic and emotional stress on drivers. Some enrolled drivers mentioned that they were required to work at designated times. Twenty-five-year-old Liu said,

I am required to work in two of the three designated rush hours: morning rush hours from 6:00am to 10:00am, afternoon from 5:00pm to 8:00pm, and late-night peak from 10:00pm to 0:00am. Satisfying the attendance requirement significantly restricted my flexibility.

Twenty-six-year-old Zhang complained that he was subjugated to Didi’s relocation requests:

I am required to maintain a minimum 85 percent of re-routing acceptance rate. this re-relocation directive essentially disfavors drivers because Didi always mobilizes more drivers than the real-time demand at will. It is too common that when I arrived at the surge zone, the surge pricing has already disappeared. However, I am not compensated for anything.

Such coercion was introduced to help Didi cope with the constant challenge of balancing demand and supply with minimal economic costs. Under the structure of flexible employment, ride-hailing platforms need to manage the supply of workers at any given time (Griesbach et al., 2019: 5). When demand surges, a typical solution is leveraging economic incentives to nudge drivers to work in certain areas or at certain times (Rosenblat, 2018; Shapiro, 2020b). However, the incentive-based behavioral nudging has inherent limitations: it is costly, and it is uncertain in many cases, as it is always possible that drivers will refuse to comply. As a response, Didi assigns the costs of keeping the platform’s labor self-sufficiency to the level of drivers, relying on forced labor to get the job done. Compared with incentives, coercion relieves Didi of the costs and uncertainties.

In addition to all these compulsory requirements, what irritated my interviewees most is the ever-decreasing remuneration. For drivers with their own vehicles, Didi often offers some rewards after finishing a certain amount of rides every day. However, some enrolled drivers revealed that Didi gradually canceled all rewards and bonuses. As 28-year-old Fu remarked: “no extra rewarded at all—it becomes my duty to serve Didi to death.” As Breman (2010) argues, when debt serves as a mechanism of attachment, the debtors have to repay the providers in labor, if and when desired, for a price lower than the existing market rate.

Leasing contracts create a legitimized situation of dependence that seriously limits the negotiating power of drivers. While workload and disciplines imposed on drivers lead to shared grievances among them, collective resistance is most unlikely to occur because of the debt obligation. Debt causes drivers to comply with worsening working conditions and decreasing wages that would be unacceptable to others. Trapped in a 3-year vehicle rental contract, my interviewee Yin must keep paying off the debt every month by working long hours under highly exploitative terms and despotic rules. When grievances are escalated, he tends to absorb them himself. What he did was to work hard in the hope of getting out of the situation sooner. As he said,

The worst thing about driving for Didi while paying for the rental is that you didn’t have alternatives. You are forced to be online all the time, forced to drive overtime, forced to take most of all the orders it dispatched, forced to accept an increasingly lower wage, and in short, it is a nightmare. What can you do? If you leave before the end of your rental contract, you need to pay for the breach of contract, and you can’t get your upfront deposit back, but they are a lot for me . . . So, it is really like a nightmare . . ..

The entrepreneurial project is a key step in developing one of the most central business strategies of Didi. Starting from 2016, Didi has launched various vehicle-leasing programs intended for drivers who do not have a car or whose car cannot meet city standards. One of the most popular projects is annual-based car rentals. The rent is dependent upon vehicle types and locations. For example, in Beijing, vehicles like Passat and Camry cost around 6000 RMB per month. To rent a car, drivers also need to put up a one-time deposit of 15,000 RMB to 20,000 RMB. Surveys from different sources demonstrate that more than half of Didi’s drivers rent vehicles to work in China’s big cities, and the percentage of rent-to-drive laborers has been increasing in the past 3 years (Qi and Li, 2020). Although they do not have a formal labor contract with Didi, they work full time, with an average week of 70 driving hours (Qi and Li, 2020). With leased vehicles, most drivers struggled with low piece rate, compulsory long working hours, despotic rules, and health hazards. The physical intensity of the labor, long hours sitting, the lack of drinking water, and irregular eating expose drivers to illnesses such as back pain, stomach pain, and kidney disease. Due to the necessity of getting back the deposit, drivers have a compromised ability to walk away. The system of forced labor was a driving force behind Didi’s emergence as a ride-hailing giant.

Some might argue that Didi’s situation is not coercive enough to qualify as forced labor—after all, there are no visible signs of physical restraint and abuse. Besides, despite some limitations, drivers are ultimately free to quit if they pay for liquidated damages and give up deposits. Unfortunately, all these arguments cannot capture the full range of forced labor relations in the capitalist economy. First, coercion is exemplified either explicitly or implicitly by a threat. Didi’s control over technologies grants the platform expansive punitive power—it can easily deny drivers’ permission to access the app or stop dispatching new orders to them. Although these sanctions are temporary, their impact is consistent because drivers are working “under the constant threat of punishment” (Hatton, 2020: 11). The feeling of potential risk produces a docile labor force. Second, the contractual lease and labor service contract serve as the legal basis that justifies Didi’s labor management and drivers’ responsibilities. Consequently, with the continuing lease, drivers are unable to refuse without facing some kinds of punishment. Finally, considering that most rent-to-work drivers are migrant workers who are in economic hardship (Qi and Li, 2020), a great many are not fortunate enough financially to exercise the freedom of walking away. The combination of multiple factors—debt, contract, and technological subordination—render Didi’s power over drivers almost absolute.

Didi’s partner–entrepreneurial projects target not only prospective drivers but also micro-entrepreneurs who engage in the business of vehicle leasing. Instead of making deals with auto companies, Didi discharges the financial burdens and risks onto FMCs. FMCs work on commission by engaging in car rentals, driver recruitment, and fleet management. For example, one FMC I interviewed managed 200 vehicles. These vehicles were acquired from different resources—very few firsthand vehicles, some secondhand ones, and most leased from auto companies. FMCs profited from brokering the relationship between Didi and drivers, earning in the range of 1–5% from drivers.

The revenue-sharing mechanism provides FMCs with incentives to impose normative and ideological control over drivers. In other words, Didi manages drivers through both technological and organizational means that involve a hierarchy of algorithms, labor brokers, and supervisors at local level. FMCs exercise direct supervision over drivers’ work on a daily basis, monitoring their performance, ensuring compliance, and urging drivers to transform labor power into productive power to the utmost (Li, 2021). An FMC manager said, “Drivers are like gas that fuels vehicles.” With this tenet, the FMC exerts additional requirements on drivers under the name of Didi. For example, one driver affiliated with this FMC mentioned that he can only have 3 days off every month. Some new drivers tend to be resistant after realizing that they were cheated by Didi or/and their FMCs. However, with an ongoing rental, drivers faced minimal options. As the manager said, “reversing the contract is impossible unless drivers pay the penalty.” Therefore, most drivers tend to comply—even if with great reservations—with the terms of work set by Didi and FMCs, so that they can just quit. Many drivers see problems as soluble by individual exit; therefore, they have the preference to endure and then leave, rather than resisting.

Political and economic logics of neo-bondage

Two factors gave important stimuli to the emergence of neo-bondage at Didi. One is an internal factor that speaks to the labor shortage. As Robert Evans rightly argues, the use of bonded labor is typically associated with the phenomenon of labor shortage, but raising wages is not considered a viable solution to this shortage (Evans, 1970: 861). This is exactly true of the situation of ride-hailing companies. As scholars suggest, while classifying drivers as independent contractors saves costs, it also leads to uncertainty and fluctuations in labor supply (Li, 2021; Shapiro, 2020b). The existing literature argues that turning to monetary incentives is a central strategy that platforms use to keep sufficient drivers. Through massive subsidy, platforms enlist a large pool of drivers; by leveraging monetary incentives, they subtly reorient drivers toward certain places at certain times. However, in the eyes of Didi, economic incentives are not considered a viable solution that can last indefinitely because they are costly. My interviewee, a Didi data scientist, commented on the monetary incentives in this way:

While incentives play a vital role in the initial market competition, Didi has consistently reduced bonuses awarded to drivers after obtaining more than 80 percent market share in China. With declining wages, some drivers quit, which can, in turn, result in the unstable and insufficient supply of drivers. Thus, Didi has been making efforts to find alternative ways to secure its driver force, because drivers are the tools through which Didi makes money.

Thus, if the crisis of labor shortage is a constant in the business model of the gig economy, forced labor becomes a strategy to create a reliable supply of workers. In the eyes of Didi, this forced labor regime finds its rationale in the objective to build a dependable workforce and simultaneously skirt the legal responsibilities associated with conventional employment relations. Feng, a former senior manager in Didi in charge of multiple vehicle-leasing programs, explicitly explained this point,

To pay off the monthly fee, drivers need to be completely committed to the work. Therefore, these programs enable Didi to achieve substantial control over drivers, preventing them from quitting and striking.

During the interview, Feng recalled how excited Didi’s managers felt when they saw a significant increase in active drivers and their driving hours. From this perspective, I emphasize that drivers were forced into neo-bondage because Didi desired bonded subjects. The neo-bondage system enables Didi to create a dependable and disposable workforce, while simultaneously minimizing costs.

While economic logic does play a significant role in shaping neo-bondage model, regulatory contexts introduced further complexities into an already uncertain system. In July 2016, China launched regulatory guidelines for the ride-hailing sector (hereafter the 2016 regulation). The 2016 regulation is characterized by its central-local dualism, with a combination of a neoliberal-style national policy and stringent and diversified local policy (Jiang and Zhang, 2019). Under the macro policy imperative of promoting the digital economy, the central state legalized the ride-hailing business. However, given the diversity of local markets, the central state shifted the regulatory power to local governments to develop additional rules (Chen and Qiu, 2019). As a result, nationwide, at least 229 cities proposed local standards, and they varied substantially. For example, some cities set specific standards on vehicles, including requirements like the minimum wheelbase at 2700 mm, under 600,000 km on the odometer, and car price higher than 120,000 RMB; others restricted ride-hailing drivers to residents with a local vehicle plate and a local permanent residency (or local hukou in Chinese) (Jiang and Zhang, 2019).

These regulatory variations had manifold effects; it is beyond the article’s scope to narrate all of them. However, the most immediate and relevant to this article is that they resulted in a sharp decline in Didi’s active drivers (Chen and Qiu, 2019). Requirements on the standards of vehicles significantly elevated the entry barrier, for example. In addition, the requirements to have local drivers and local plates almost put Didi’s whole business at risk because its drivers are primarily made up of migrant workers who do not have local residency and/or plates (Chen and Qiu, 2019).

While the 2016 regulation was formulated in the spirit of normalizing the ride-hailing industry, it produced somewhat counterproductive impacts. Instead of making Didi sign up qualified drivers in accordance with the requirements, it pushed Didi to operate in a gray zone. Vehicle-leasing programs emerged as a response to the emerging regulatory threats. Through these programs, Didi can at least ensure vehicles—if not drivers—operating on the platform meet the cities’ standards.

It was also imparted to drivers —despite a lack of concreteness—that signing vehicle-leasing contracts with FMCs would guarantee their legal status (Qi and Li, 2020), because the leased cars will be registered as for-hire vehicles and drivers can obtain a ride-hailing license, both of which are required by legal departments. Combined together, the economic and regulatory environment enables us to contextualize Didi’s vehicle-leasing initiatives.

Discussion and conclusion

The purpose of this study is to gain a better understanding of the important roles of labor instruments in the ride-hailing economy. Whereas prior research suggests that ride-hailing platforms tap into existing and under-utilized resources to expand, my article shows that Didi produced a fresh market of vehicle leasing that provides potential drivers with the means of production on incurring debt. To further conceptualize the debt-based forms of production and labor relations, I draw upon the neo-bondage concept that essentially explains the persistence and renewal of bonded labor in the contemporary world. I argue that in an environment characterized by a labor shortage and regulatory uncertainties, Didi leveraged neo-bondage to secure a dependable and disposable labor force while externalizing regulatory threats. Debt serves as the leading edge of coercion that subjects drivers to providing services on an involuntary basis. This article highlights three key indicators of Didi’s forced labor practices—deception and fraud in the recruiting stage, compulsory requirements in the labor process that drivers are forced to follow, and decreasing remuneration that drivers have to accept. This article also explains how Didi outsourced the business of leasing vehicles to the FMCs. I argue that this arrangement has produced organizational control that further restrains drivers’ freedom and spaces for collective action.

This article has two theoretical implications. First, by bringing the largely neglected labor instruments into consideration, this article challenges the still-influential discourse of the sharing economy (Rosenblat, 2018; Schor, 2021; Shapiro, 2020a). To some extent, this sharing language successfully conceals the debt-based forms of organization and production. While vehicle-leasing programs prevail in many other places, such as Uber and Lyft in the United States and Grab in Singapore, they are largely neglected by the existing literature. Therefore, the findings of this research could lay the solid foundation for further research on the topic and comparative studies.

Second, debt is invariably associated with some forms of labor coercion. Thus, this article challenges the scholarly focus on the issue of precarity in the gig economy (Hyman, 2018; Prassl, 2018; Ravenelle, 2019) and calls scholarly attention to forced labor issues. As widely documented in the existing literature, forced labor is not a remnant of the past but a key feature of present-day capitalism. Media scholars have analyzed how factory workers on Foxconn’s electronic assembly lines suffer from slave-like working conditions characterized by compulsion and violent threats (Ngai and Chan, 2012). This article highlights the renewal of forced labor in the ride-hailing industry, although the manifestations of this forced labor are more hidden and cryptic. Moreover, it is worth emphasizing that labor coercion is not a phenomenon only found in Didi but is distributed more broadly across different sections of China’s growing platform economy. For example, sociologist Lei Yawen in her work on China’s food-delivery platforms argued that “the platform-courier relationship is a master-slave relationship” where couriers are forced to take orders at low pay (Lei, 2021: 299). Thus, the findings in the article have broader implications for research into other gig platforms in China.

Finally, it is important to acknowledge two limitations of this research. First, scholarly publications move more slowly than large and rich Internet companies. Therefore, in this article, I have not attempted to keep up with the company’s most recent initiatives. Instead, I have tried to articulate broad and significant themes and patterns that should hold constant for some years and deserve further scholarly research. Second, to be sure, a broad understanding of forced labor not only considers it as an economic relationship but also concerns the political, institutional, and ideological relations wherein the forced labor regime is sustained. That said, employers’ punitive power is bolstered by legal frameworks, the relations of job positions to the labor market, room for labor movements to organize, and so on (Lei, 2021). Nevertheless, rather than simultaneously analyzing all the factors and processes, this study pays more attention to the grounded evidence. However, the broader historical and social background that contextualizes the development of the forced labor regime deserves further research.

Footnotes

Acknowledgements

I gratefully thank Amanda Ciafone, Clifford Christians, and Jeffery Martin for their sustained support and mentorship. I thank Renyi Hong, Yawen Lei, and Hao Qi for their insightful comments on my ideas on various occasions. In addition, the article also benefited from an invited talk at East Asian Institute (EAI) at NUS, where Emily Chua Hui Ching and Jiwei Qian offered constructive suggestions. I thank two anonymous reviewers and the editor for their excellent suggestions and encouragement. Finally, many thanks go to my informants and interviewees who made my research possible.

Author’s Note

The author has agreed to the submission of the article and that the article is not currently being considered for publication by any other print or electronic journal.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The author(s) received financial support from Chiang Ching-Kuo Foundation for International Scholarly Exchange [Grant Number DD039-A-18] and NUS Startup Grant.

ORCID iD

Angela Ke Li

Author biography

Angela Ke Li is an Assistant Professor at the Department of Communications and New Media, National University of Singapore. She gained her PhD from the Institute of Communication Research, University of Illinois, Urbana-Champaign. Her research and teaching are at the nexus of various fields, including political economy of communication and culture, science and technology studies, and anthropology to contemporary China. Previous work appears in journals including Information, Communication & Society, International Journal of Communication, Journalism, Journalism Studies, China Perspectives, Chinese Journal of Communication, and Global Media and China.

References

Acquier

(2018) Uberization meets organizational theory: platform capitalism and the rebirth of the putting-out system. In: Davision

Finck

Infranca

(eds) Cambridge Handbook on Law and Regulation of the Sharing Economy. Oxford: Cambridge University Press, pp. 13–26.

Baishya

(1997) The putting out system in ancient India. Social Scientist 25(7/8): 51–56.

Breman

(2010) Neo-bondage: a fieldwork-based account. International Labor and Working-Class History 78: 48–62.

Breman

(2014) On labour bondage. Contributions to Indian Sociology 48(1): 133–141.

Calvão

(2016) Unfree labor. Annual Review of Anthropology 45(1): 451–467.

Chan

(2019) Becoming an expert in driving for Uber”: Uber driver/bloggers’ performance of expertise and self-presentation on YouTube. New Media & Society 21(9): 2048–2067.

Chan

Kwok

(2021) The politics of platform power in surveillance capitalism: a comparative case study of ride-hailing platforms in China and the United States. Global Media and China. Epub ahead of print 26 September. DOI: 10.1177/20594364211046769

Chen

(2021) Didi raises $4.4 billion in largest China-US IPO since Alibaba. Technode. Available at: https://technode.com/2021/07/01/didi-raises-4-4-billion-in-largest-china-us-ipo-since-alibaba/ (accessed 5 November 2021)

Chen

(2018) Thrown under the bus and outrunning it! the logic of Didi and taxi drivers’ labour and activism in the on-demand economy. New Media & Society 20(8): 2691–2711.

10.

Chen

Qiu

(2019) Digital utility: datafication, regulation, labor, and DiDi’s platformization of urban transport in China. Chinese Journal of Communication 12(3): 274–289.

11.

Evans

(1970) Some notes on coerced Labor. The Journal of Economic History 30(4): 861–866.

12.

Friedman

Lee

(2010) Remaking the world of Chinese labor: a 30-year retrospective. British Journal of Industrial Relations 48(3): 507–533.

13.

Griesbach

Reich

Elliott-Negri

, et al. (2019) Algorithmic control in platform food delivery work. Socius: Sociological Research for a Dynamic World 5: 1–15.

14.

Guérin

Venkatasubramanian

Kumar

(2015) Debt bondage and the tricks of capital. Economic and Political Weekly 50(26/27): 11–18.

15.

Hatton

(2019) “Either you do it or you’re going to the box”: coerced labor in contemporary America. Critical Sociology 45(6): 907–920.

16.

Hatton

(2020) Coerced: Work Under Threat of Punishment. Oakland, CA: University of California Press.

17.

Hyman

(2018) Temp: How American Work, American Business, and the American Dream Became Temporary. New York: Viking.

18.

Jiang

Zhang

(2019) An experimental model of regulating the sharing economy in China: the case of online car hailing. Computer Law & Security Review 35(2): 145–156.

19.

Kara

(2008) Sex Trafficking: Inside the Business of Modern Slavery. New York: Columbia University Press.

20.

Kenney

Zysman

(2016) The rise of the platform economy. Issues in Science and Technology 32(3): 61–69.

21.

Kumari

(2018) Neo-bondage in the brick kiln industry: a case study of Bihar. Social Change 48(3): 384–397.

22.

LeBaron

(2014) Reconceptualizing debt bondage: debt as a class-based form of labor discipline. Critical Sociology 40(5): 763–780.

23.

Lei

(2021) Delivering solidarity: platform architecture and collective contention in China’s platform economy. American Sociological Review 86(2): 279–309.

24.

(2021) Beyond algorithmic control: flexibility, intermediaries, and paradox in the on-demand economy. Information, Communication & Society. Epub ahead of print 23 May. DOI: 10.1080/1369118X.2021.1924225.

25.

McGrath

(2013) Many chains to break: the multi-dimensional concept of slave labor in Brazil. Antipode 45(4): 1005–1028.

26.

Marx

(2014) Economic and Philosophic Manuscripts of 1844. New York: Dover Publications Inc.

27.

Ngai

Chan

(2012) Global capital, the state, and Chinese workers: the Foxconn experience. Modern China 38(4): 383–410.

28.

Prassl

(2018) Humans as a Service: The Promise and Perils of Work in the Gig Economy. London: Oxford University Press.

29.

(2020) Putting precarity back to production: a case study of Didi Kuaiche drivers in the city of Nanjing, china. Review of Radical Political Economics 52(3): 506–522.

30.

Ravenelle

(2019) Hustle and Gig. Oakland, CA: University of California Press.

31.

Rosenblat

(2018) Uberland: How Algorithms Are Rewriting the Rules of Work. Oakland, CA: University of California Press.

32.

Schor

(2021) After the Gig: How the Sharing Economy Got Hijacked and How to Win It Back. Oakland, CA: University of California Press.

33.

Shapiro

(2020a) Design, Control, Predict: Logistical Governance in the Smart City. Minneapolis, MN: University of Minnesota Press.

34.

Shapiro

(2020b) Dynamic exploits: calculative asymmetries in the on-demand economy. New Technology, Work and Employment 35(2): 162–177.

35.

Shields

(2011) Labor exploitation. In: Kaufmann

Kuch

Neuhäuser

, et al. (eds) Humiliation, Degradation, Dehumanization: Human Dignity Violated. New York: Springer, pp. 173–191.

36.

Skrivankova

(2010) Between decent work and forced labour: examining the continuum of exploitation. JRF Programme Paper. Available at: https://www.jrf.org.uk/report/between-decent-work-and-forced-labour-examining-continuum-exploitation (accessed 15 March 2021)

37.

Stanford

(2017) The resurgence of gig work: historical and theoretical perspectives. The Economic and Labor Relations Review 28(3): 382–401.

38.

Strauss

McGrath

(2017) Temporary migration, precarious employment and unfree labour relations: exploring the “continuum of exploitation” in Canada’s temporary foreign worker program. Geoforum 78: 199–208.

39.

Sundararajan

(2017) The Sharing Economy: The End of Employment and the Rise of Crowd-based Capitalism. Cambridge, MA: The MIT Press.

40.

Tencent Technology (2016) Didi promotes partner-entrepreneurial project, driving for Didi without a car. Available at: https://tech.qq.com/a/20160418/049987.htm (accessed 10 July 2021)

41.

Tracy

(2019) Qualitative Research Methods: Collecting Evidence, Crafting Analysis, Communicating Impact. Hoboken, NJ: Wiley-Blackwell.

42.

Zhu

Iansiti

(2019) Why some platforms thrive and others don’t. Harvard Business Review. Available at: https://hbr.org/2019/01/why-some-platforms-thrive-and-others-dont (accessed 5 Nov 2021)