Editor's note: Morshed Mannan is a Meijers PhD candidate at the Company Law department of Leiden Law School. He received his LL.M. Advanced Studies in International Civil and Commercial Law (cum laude) from Leiden University and has previously worked as a lawyer and lecturer in Dhaka, Bangladesh. Raam Dutia is currently an intern with the Doing Business Right team at the Asser Institute. He recently received his LL.M. Advanced Studies in Public International Law (cum laude) from Leiden University and has worked at an international law firm in London on a range of debt capital markets transactions.
For many, Uber epitomises the "move fast and break things" ethos of successful Silicon Valley start-ups. The company enters new markets before regulators are ready, capitalising on regulatory bottlenecks and uncertainties in numerous jurisdictions – only to enlist its enthusiastic customer base and other means to challenge regulators when they catch up. The backlash against this mode of operation has been severe, and boycotts and a litany of lawsuits appear to have dented Uber's image and plunged the company into crisis.[1] Elisa Chiaro’s recent blogpost discussed the implications of platform economy enterprises, such as Uber, on the rights and protections of workers. In this, the first of a series of blogposts, we will take a broader view by exploring whether the company’s concerted efforts to conduct operations in a way that avoids or attempts to undermine local, state and national regulations shapes the law across the markets in which it operates. This will be done by appraising the growing literature on the effect of its regulatory arbitrage[2] and evaluating whether the company’s use of algorithms, in conjunction with standardized service agreements, rider agreements and other contracts to govern the relationships between various stakeholders, establishes it as a source of transnational lawmaking within a large network of well-defined stakeholders: drivers, riders and civil society. Uber’s business practices and litigation in the UK will be used as a case study that is illustrative of broader trends. By doing so, we hope to contribute a deeper understanding of the patterns that have emerged through Uber’s local activities in several jurisdictions. In later entries, we will examine the response to these attempts at regulatory arbitrage and private ordering as well as the repercussions this has on the contemporary regulation of the platform economy.
Exploiting regulatory uncertainties
The idea of Uber was born out of solving a need. The apocryphal story of Uber’s founders waiting in the snow for a Parisian taxi that never arrived can be interpreted in at least two ways. First, as a commentary on the tired state of local transport, a system that is perceived to be slow, infrequent, expensive and inconvenient. Second, as an indictment of the regulatory framework that distorted the supply and demand of transportation services by, for instance, restricting the number of taxi medallions available in a metropolitan area at a certain time. Instead of suggesting that these inefficiencies be ameliorated by revisiting local or national legislation, the founders of Uber created a technological means to bypass regulation altogether. From the time it was known as UberCab in San Francisco to the early years of its global expansion, the company’s modus operandi has been “ride first, ask questions later”. It would roll out its services in a new market, voraciously recruit drivers and match them to passengers at scale irrespective of local licensing requirements. By the time regulators catch up, requiring Uber to obtain a private hire vehicle (PHV) operator license or banning it outright, it would have attracted a loyal base of riders and drivers with which such a measure would be deeply unpopular. While instituting measures such as aggressive litigation[3] or lobbying national or regional authorities to pre-empt municipal regulation categorising Uber as a taxi service,[4] the company leverages its popularity in order to challenge cease-and-desist orders, unfavourable ordinances and license cancellations.[5] Whether Uber's actions are undesirable remains the subject of heated debate,[6] yet the company has ultimately been able to exploit the uncertainty over its position to its advantage, helping it (at least partially) reach a valuation of around USD 48 billion in late 2017 and establishing a presence in 633 cities in over 80 countries and territories. In London alone, there are over 30,000 drivers and 2 million registered riders. As one academic persuasively argues, Uber is the archetype of the ‘postindustrial’ corporation. In lieu of productive enterprise, the company maximizes shareholder value through regulatory arbitrage and evasion, along with (ideational) asset manipulation and speculative activity.[7]
This is manifested in Uber’s bold claims to be a mere licensor of software.[8] This pretension, based on the rupture between the legal structure and economic reality of the business, has been the central cause of myriad legal dilemmas for its drivers and riders. As a purported consumer of Uber’s software, drivers are presented as independent contractors, with attendant consequences on their capacity to access worker rights, their insurance liability and their tax status. As elaborated in Chiaro’s blogpost, the misclassification of the employment status of drivers is the most litigated cause of action involving Uber. A common theme in these lawsuits is that the extent of control and supervision exercised by Uber effectively subrogates drivers as either employees or ‘limb (b)’ workers, in countries such as the UK that recognise a third employment category. This is done while denying drivers basic rights to a minimum wage and working time protections. This additionally causes uncertainty about the right of drivers to unionise or whether they are potentially engaged in an illegal price-fixing conspiracy via the Uber app. Vanessa Katz highlights how the blurring of the distinction between commercial and personal activity in the US has had implications for insurance liability, as drivers are unable to rely on personal insurance policies while driving with the app on. Uber has historically been reticent to provide commercial insurance during the entire period of the app's usage by the driver, even contesting attempts to introduce statutory insurance requirements in California.[9] In the UK, Uber drivers are required to cover their own private hire insurance. Similarly, the erosion of the distinction between personal and commercial activity through the categorisation of Uber drivers as ‘microbusinesses’ has raised tax issues given the ambiguity regarding when business deductions would be available and the difficulties in ensuring compliance and enforcement.[10] The burden of income tax filing is shifted entirely onto the driver, instead of being deducted at source by Uber.
While some of these issues are endemic in sectors where (bogus) self-employment is rife, others are particular to the operation of Uber’s app and impact both drivers and riders. For instance, one field investigation reveals that the company’s rider-sourced rating system to evaluate, monitor and sanction drivers is vulnerable to consumer bias and creates potential for employment discrimination. Conversely, another empirical study of 1,500 rides hailed on controlled routes in Seattle and Boston found a pattern of discrimination against ethnic minority riders, with African Americans facing longer waiting times and driver cancellations. Their private method of training, monitoring and evaluation has had a poor track record in ensuring rider safety, with the rape of a female passenger in India in 2014 gaining notoriety. As drivers are paid per ride and can be suspended from the app for turning down too many ride requests,[11] they are financially incentivized to accept difficult or intoxicated passengers, thereby exposing themselves to assault. Serious concerns have also been raised about how the company handles the data of its consumers, with Uber entering into a consent order with the US Federal Trade Commission in August 2017 requiring remedial action, only then to be found in November 2017 to be concealing a hack of the data of 57 million users, including the names, email addresses, phone numbers and license plate numbers of over 600,000 drivers.
While there have been an unfortunate number of settlements in disputes involving Uber, when given the opportunity courts have not shied away from providing clarity as to the substantive nature of the company’s activities. In determining whether Uber is a digital ‘information society’ service or a transportation provider, the Court of Justice of the European was forthright in its recent preliminary ruling that Uber provides transport services instead of simply digital intermediation services between a non-professional driver and a passenger using the app (para. 37). As such, they do not benefit from the more limited regulation and protection from liability to which a digital intermediary would be subject under EU law. A result of this decision can already be seen in the UK, where actions against Uber and HMRC for the establishment of Uber's Value Added Tax are being pursued with renewed vigour. Courts have been especially scathing about the contractual contortions that “armies of lawyers” have attempted so as to misrepresent the true rights and obligations of concerned parties.
Thus far, we have canvassed how the operation of Uber’s app has helped it circumvent the reach of local and national regulations, while creating an ecosystem that steers the behaviour of certain actors across the numerous locations in which it operates its platform. The following section will take a deeper look at the contractual architecture created by the aforementioned army of lawyers and the ‘choice architecture’[12] presented by its app to argue that it collectively acts as a transnational form of lawmaking, applicable to its network of drivers, riders and civil society (primarily the media, unions and other non-governmental organisations representing drivers).
Shaping the operation of law and regulation through Uber's own contractual "rules" and algorithms
To appreciate the transnational dimension of Uber’s contractual and algorithmic management, it is necessary to have a clearer understanding of Uber’s corporate structure. This structure is somewhat opaque since Uber’s global business is carried out through a network of private entities and as such do not have public disclosure obligations as extensive as companies with publicly-traded shares. The following brief description draws from the company disclosure documents that are currently available and an informative report by Fortune in 2015 on Uber’s tax structure.
Uber Technologies Inc., the parent company, is registered as a Delaware Corporation and has its headquarters in San Francisco. However, its non-US operations are carried out from the Netherlands, where it has incorporated 10 subsidiaries. The top subsidiary is Uber International C.V., a limited partnership that is registered in Amsterdam but headquartered in Bermuda, which pays the parent royalties based on net global revenues, in exchange for the right to use the intellectual property developed by the parent. Uber International C.V. has another IP agreement with Uber B.V, one of the other Dutch subsidiaries, that permits the latter to use the application in exchange for royalties that amount to 99% of net revenue. This technology is then deployed by Uber B.V. worldwide. Uber B.V. in turn receives each payment made by riders to drivers through the application, deducts a certain percentage for its own income and via Rasier Operations B.V., another Dutch limited liability company, processes the payment made to drivers. While Uber has subsidiaries in the countries it operates, usually in the form of a private limited company, this is generally for recruiting drivers as well as marketing and support services, such as lobbying for favourable regulation and protecting brand image. In addition, Uber has a regional office in Singapore and has opened an engineering centre in Bengaluru, India. It would seem, then, that Uber's business strategy is transnationally directed and overseen from its two global hubs in California and the Netherlands.
Transnational lawmaking through contract: the UK as case study
These entities enter into a web of contracts with drivers and riders across the globe. Uber's use of contract in respect of its UK operations provides a representative illustration of this. Transactions are conducted by way of a triangular contractual relationship, with a Services Agreement entered into between Uber B.V. (UBV) and the driver, on the one hand, and a Rider Agreement between a separate Uber entity holding an operator's license, Uber London Ltd. (ULL), and the passenger, on the other. The language in the Rider Agreement makes clear that ULL provides booking services as an "agent" (Part 1, clause 1) to the driver/transportation provider (not party to the agreement) and acts also as a payment collection agent to the driver as a principal (Part 2, clause 4). The Services Agreement appears consistent with this wording. UBV's role, again, is that of "payment collection agent" as well as providing support and "on-demand intermediary and related services" to the driver (clause 1.17). Moreover, the Services Agreement refers to the driver as an "independent company in the business of providing Transportation Services" and an "independent contractor" (clause 13.1). The Services Agreement also purports to create a "legal and direct business relationship" between the driver and passenger (clause 2.3), notwithstanding the lack of any clear contractual relationship between the passenger and driver. The Services Agreement also states that "Uber [UBV] and… [ULL] do not, and shall not be deemed to, direct or control" drivers or their performance under the agreement (clause 2.4). As we have seen, as a consequence of this characterisation of the parties’ triangular relationship, protections under employment law were elided and obligations under licensing, insurance, competition and tax were added for drivers. In turn, this claim to act as a technological agent eased the regulatory burden on Uber, particularly in relation to Value Added Tax.
While there were extant laws to address such situations in the UK, they were not enforced. Between its launch (July 2012) till the act of state enforcement (i.e the decision of the UK Employment Tribunal in Aslam, Farrar and Others in October 2016), the company carved out a space in which contract created certainty as to the rights and responsibilities of Uber, its drivers and riders.
Supporting the contractual architecture through algorithmic management
The control and direction that Uber exerts through this triangular contractual arrangement has been buttressed by the workings of its app, largely replicated across its global network. The uniformity of its appearance and seamlessness of its functioning is integral to the app’s appeal. The user interface is practically identical, whether used in India or the UK. Once a user downloads the passenger app and creates a profile, they are able to request a ride through one of Uber's service options. When a nearby driver confirms their availability for the trip on the driver app - being induced to do so at risk of being automatically logged off the app or suspended for rejecting too many trips - it confirms the ride and provides the driver's first name and license plate number to the rider. Once the rider has been picked up, drivers are prompted to take the route suggested by Uber, having to justify any deviations in case of undue delay. Once the ride is over, Uber charges the fare, the base rate being set by Uber's algorithms and from which a driver may only negotiate downwards, to the rider's credit or debit card and remits the amount to the driver less Uber's service fee. Crucially, Uber operates and monitors a feedback mechanism, whereby the rider and driver rate each other on a scale of 1 to 5. Drivers unable to improve poor scores are called in for quality checks and risk having their accounts deactivated. The supply and demand of trips are also manipulated by Uber’s surge pricing algorithm, with heat maps indicating that fares for certain undersupplied areas are temporarily increasing to entice drivers.[13] Through this system, Uber is able to nudge, control and algorithmically manage its drivers, incentivise good behaviour and engender trust between strangers.[14] While drivers can ostensibly log-off the app at any time, the combination of high initial investment (e.g. acquiring PHV license and insurance; acquiring a model of car that Uber prefers) and financial reliance on the company acts as a strong deterrent.
This cumulative set of practices is known as algorithmic management. It allows a handful of managers to direct and supervise thousands of drivers. What makes it so effective is the large volumes of data that it gathers on travel times, weather, driver and rider movement, their characteristics and their preferences. The data, gathered every 4 seconds, is stored on the drivers’ phones and in data centres across six continents. The company has two data controllers, Uber Technologies, Inc. for US residents and Uber B.V. for non-US residents, who determine the purposes for which data is processed. This data is used to help Uber’s algorithms learn where drivers and riders are, forecast where they are expected to be and improve routing from point A to B. While this is used to enhance the service provided by individual drivers at a local level, this data is also utilised at the global level to, for instance, visualise demand patterns across cities and thereby further improve the Uber app. As we can see, the big data generated through the operation of its algorithms aids Uber in both its local and global strategies.
Moreover, Uber appears to have shaped the operation and enforcement of the law in an altogether more blunt way through a worldwide programme to deceive authorities in markets where its services are being resisted or have been altogether banned. Following a NY Times investigation, it was revealed that through a process known as "Greyballing", blacklisted enforcement officials have been identified and prevented from hailing rides by showing a bogus map of an area and seemingly cancelling (non-existent) rides before they arrive. This has hampered the ability of local governments to enforce their own regulations.
We contend that this web of contracts, in conjunction with algorithmic management has rule-making effects, with Uber proffering its own interpretation of local and national regulations, which it then applies and even enforces itself. In so doing, we argue that the company has come to define the behavioural framework in which Uber, its numerous drivers, riders and civil society interact. Paraphrasing Backer (2008), who has argued that multinational enterprises such as Walmart and Gap are sources (and objects) of transnational regulation, this framework can be seen as a “freestanding… self-communicating [regulatory] system” (p. 517) that interacts (and develops in concert) with traditional public law. In his research, Backer identifies how a network is built between actors (e.g. a company, contractors, NGOs, media, consumers) through repeated interactions, instigated by a company at the nexus of this global system. The company issues a set of standards that it embeds in its supplier contracts, compliance of which is ensured through private monitoring and sanctions, including remedial action. These contracts function everyday, governing innumerable transactions over dozens of countries, with the standards acting as de facto labour codes. If there is an egregious violation of this private regulation or substantive law, for instance through the usage of child labour, organisations that form civil society, such as the media, trade unions and human rights NGOs act to reveal and challenge these failures. Consumers – and increasingly banks and institutional investors – are thus invited to reconsider their relationship with the company, exercising significant financial pressure. While acknowledging the ontological claim that the nature of law involves both the state and private arrangements,[15] the states across which these multinational enterprises operate are present but often take a backseat to transnational private regulation, either due to the economic costs of enforcement or the difficulties in enforcing national law across borders. Whether Uber operates a parallel regulatory system to quite the same degree is yet to be determined, but the company can be regarded as overlaying or interacting with traditional legal frameworks with its own claims as to what the law is or should be, in effect transforming its application.
Throughout this blogpost, we have demonstrated how Uber’s contractual and algorithmic practices in modifying behaviour indicate a similar – though more subtle – pattern of transnational private lawmaking, thereby producing “effects on social welfare similar to the effects resulting from rule-making and enforcement by governments”.[16] We have shown how significant violations of this regime, for instance through a massive data breach, elicits a strong response from the media and privacy groups. These civil society groups, including unions and organisations that represent drivers, have been at the forefront of exposés regarding Uber’s activities. This has evoked a swifter reaction from Uber, in terms of addressing its data protection practices (e.g. removing its chief of security), than the coordinated activities of national data protection authorities. While Uber’s activities only affect a narrow segment of economic actors, they are tied together by economic and social interest. Thus, Uber has, at least until the most recent spate of litigation and regulation, managed to define and set the rules for its own operations as a form of transnational regulation which is enmeshed in the local and the national. At the very least, Uber has blurred the relationship between itself and regulators transnationally by interpreting the scope of the existing regulatory framework and, crucially, imposing its interpretation of these regulations over numerous actors, both public and private. It effectively defines, ex ante, how its activities are to be characterised under the law and how the relevant regulations are to apply to itself.
The next blog will take stock of recent legal developments, noting the extent to which Uber is an object of, and thus shaped by, regulations across the globe. It will examine the extent to which legal systems have been able to resist Uber's attempts to circumvent its operation and new avenues of engagement between the controversial platform and regulators.
[1] Not to mention corporate governance concerns raised in 2017, as evidenced by the exposure of sexual harassment suffered by its employees, allegations of having stolen trade secrets and bribed foreign officials and the rapid departure of its co-founder and CEO Travis Kalanick in mid-2017.