Abstract
Multiparty constellations are on the rise in the labour market, and they can make the classification of contractual relationships exceedingly difficult. Recent case law on platform work provides an insight into the various problems resulting from this development. The article provides an overview of cases in which courts and/or administrative bodies across Europe were called upon to rule on platform workers’ rights in cases that involved relevant multi-party constellations. It establishes a typology of the different actors that have figured as defendants in cases on platform workers’ rights and explores the consequences for both procedural and material aspects. On the basis of these insights from case law, as well as a brief review of the European Parliament's suggestion of regulating subcontractor liability in the proposed directive on platform work, a number of regulatory implications are identified.
Keywords
Introduction
As the classification of certain types of platform work as employment relationships in various jurisdictions across Europe is becoming an unmistakable reality, platform models are evolving, and with them the legal questions that arise and may ultimately call for a resolution in court. One aspect of this evolution is that questions of the ‘if’ of the existence of an employment relationship may increasingly need to be complemented by questions of the ‘with whom’. This apparently trivial question can be complicated in various ways, 1 some of which have already made their way into the growing body of litigation on platform workers’ labour rights.
Many of the most well-known platforms have long been using different models for staff acquisition, and are thereby not or no longer limiting themselves to the direct contracting with individual workers, who have typically been prevented from or restricted in using replacements or substitutes without the platform's permission. 2 Platforms’ motives for exploring other options may be diverse, including e.g., the incompatibility of a ride-hailing platform's original model with local taxi regulations, but certainly also concerns that solo self-employed may be reclassified as direct employees of the platform – as has happened in one form or another in 14 European countries. 3 Notably where removing the elements that indicate subordinate employment would be difficult with a view to the platform's business model, a viable alternative might be to divert those elements to the platform worker's relationship with either the client or a subcontractor, rather than the platform itself. Additionally, some of the cases to be discussed below also indicate that working with subcontractors that oblige (rather than just incentivise) employees to work at times of high demand can be a strategy for a platform to gain a competitive edge over its competitors. 4 Finally, as platforms determined to present their workers as genuine self-employed collaborators tend to supplement their terms and conditions with more permissive substitution clauses, multiparty relationships can in principle also be formed bottom-up, by a platform worker's individual decision to subcontract (parts of) their tasks to another individual.
In light of these developments, it seems about time that the current policy debate on platform workers’ labour rights be complemented by an awareness that the relationships at issue cannot necessarily be adequately assessed when departing from a bipartite relationship. The at times virtually absurd complexity of the involvement of players can best be illustrated by the numerous lawsuits brought before the Spanish courts by drivers having worked for ride-hailing platforms. In one case, 5 a group of 12 drivers brought proceedings against a total of 16 defendants – claiming that platform Cabify was circumventing the rules on collective redundancies by having drivers hired via a multitude of subcontractors, which then proceeded to dismiss them or not prolong their contracts at roughly the same time. And while the court did not allow this lawsuit to proceed (as it found that for ruling on the preliminary procedural question of whether the plaintiffs’ cases could be joined it would have needed to anticipate the outcome of the substantive questions), numerous others were admitted against a multitude of defendants, 6 which included platform companies (with or without their parent companies and/or strategic business partners), subcontractors, temporary work agencies, liquidators and guarantee institutions in the sense of Art. 3 of the Employer Insolvency Directive 2008/94/EC (EID).
To facilitate the taking on bord of this dimension of the present day labour platform economy in the academic and regulatory discussion, this article will give an insight into cases in which courts and/or administrative bodies were called upon to rule on platform workers’ rights in cases that involved relevant multi-party constellations. To this end, it will provide an overview of cases in which more than one party has been claimed or found to have played the role of an employer or a user (in a temporary agency relationship) of a platform worker. More specifically, section 2 will give an overview of the different types of actors that have figured as defendants in cases on platform workers’ rights and explore the consequences of complex multiparty situations; section 3 discusses the degree to which the identity of the employer matters for the enforcement of such rights; section 4 presents and briefly discusses the European Parliament's (EP's) suggestion of regulating subcontractor liability in the proposed directive on platform work; and section 5 assesses the regulatory implications and concludes.
Who is the employer? Taking stock of case law across Europe
This section gives a structured overview of cases ruled on by courts or administrative bodies in Europe, in which the platform worker's employee status was either claimed by the plaintiff or common grounds between the parties, or established by the decision-making body, and in which not (only) ‘the platform company’ appeared as a defendant against whom labour rights were claimed. For this purpose, it first enlists and explores the main categories of defendants involved in the cases so far, before focusing on those proceedings in which judges were concretely asked to decide between a platform and its subcontractor(s) as to who had exercised the key functions that required their classification as employer (or in any event as a user undertaking in a temporary agency work relationship).
Defendants in cases on labour rights: the whole variety
All of the following have been identified in the existing cases as (potentially) liable for platform workers’ entitlements:
Constituents of a platform's (multinational) business structure
To begin with, a digital labour platform is rarely one company. The most important players are generally multinationals.
There does not seem to be a single case in which European platform workers have attempted to hold a non-European parent company of a platform business (such as Uber Technologies, Inc., based in San Francisco) liable for entitlements based on an employment contract. However, a number of platforms operate as groups with European headquarters and subsidiaries in all European countries where they operate, whereby it is frequently less than obvious where the decisions that determine work organisation via the platform – and thus the working conditions of an individual – are actually taken. Several of these platforms have already been involved in labour rights claims outside the country where their headquarters are located – including Uber 7 and Just Eat 8 (headquartered in the Netherlands), Deliveroo 9 (headquartered in the UK), Delivery Hero 10 (Germany), Take Eat Easy 11 (Belgium), Stuart 12 (France), Glovo 13 (Spain), Wolt 14 (Finland), Bolt 15 (Estonia) and Amazon 16 (Luxembourg).
In this regard, claimants across countries have so far generally tended to identify (one of) the national subsidiaries as defendants when arguing that the platform had performed employer functions. Uber BV appears to constitute somewhat of an exception, given that the Dutch parent company assumed the role of defendant – alone or alongside one or more of its national subsidiaries – in most of the seminal Uber cases across countries. This includes countries such as Belgium and France, where e.g., the Deliveroo cases were exclusively taken against the respective national subsidiaries, 17 or the UK, where only Stuart's national subsidiary was (successfully) claimed to be an employer. 18
Uber is also the subject of the seemingly only higher-instance rulings that concern exclusively the distinction between parent company and subsidiary. While all the Swiss courts and institutions involved agreed that Uber should be seen as an employer for purposes of social security law, 19 it took two second-instance rulings 20 and one last-instance judgment 21 to decide that contributions were due exclusively by Uber BV. All these decisions see the Dutch parent company as the sole actor configuring all aspects of the platform that resulted in the worker's subordination, so that Uber Switzerland could not be held liable. The French courts have tended to take a similar approach in those cases where the issue is addressed directly (which is usually not the case in the countless lawsuits against Uber in France). 22 These decisions 23 found the national subsidiary's role to be largely reduced to marketing activities, leading them to rule that there is an employment relationship between the plaintiffs and Uber BV, but to absolve Uber France SAS and sometimes also other group members (such as Uber Management BV or Uber Partner Support France SAS) 24 from any responsibility – with one of these rulings being confirmed by the Cassation Court. 25 By contrast, a recent first-instance ruling 26 emphasises the collaboration of all group members, e.g. whereby Uber France SAS is in charge of verifying the drivers’ identity and documents and creating their profile, leading to a confirmation of liability of ‘both Ubers’ in solidum for the worker's claims. While not addressing the point, the second Cassation Court ruling in favour of Uber's employer status 27 also appears to relate to both defendants. However, in the UK, where the plaintiffs had identified three Ubers as defendants, the courts chose one of the national subsidiaries (Uber London Ltd) as the entity with whom the drivers had an (unwritten) contractual relationship as ‘(limb (b)) workers’ when working via Uber BV's app. 28
Perhaps the most illustrative discussion of the question is found in an Italian first-instance decision 29 (later upheld on appeal) 30 which explains (author’s translation): ‘The sole shareholder of Uber Italy S.r.l. is Uber International Holding B.V., based in Amsterdam […]. It has emerged from the documents that there are further companies in the so-called Uber galaxy that interact closely with Uber Italy S.r.l., in particular Uber Portier B.V., which deals with the signing of contracts and formally with the management of the digital platform used by restaurants called Uber Eats, as well as of the application that the cyclists were forced to download, [and] Uber B.V., which assists Uber Portier in its business.’ The court saw a ‘mixing of tasks between the various companies, all engaged to carry out activities aimed at achieving the corporate purpose, as well as full decision-making powers of the defendant, Uber Italy S.r.l., in the management of the activity carried out by the cyclists’ and accordingly considered all indications of authority over the workers due either to individual decisions of Uber Italy's management or the platform's functioning as attributable to Uber Italy S.r.l., resulting in a confirmation of employer status. 31
Uber's competitor Cabify, whose European activity has so far been limited to Spain, in turn presents an illustrative example of the complexity that a platform's company structure may have even within one country. In one among a multitude of cases brought primarily against formally entirely independent subcontractors, the Catalonia Appeals Court 32 introduced the platform as follows (author’s translation): ‘CABIFY ESPAÑA S.A. (formerly MAXI MOBILITY SPAIN S.L. […] The owner of all its shares is CABIFYMOBILITY INTERNATIONAL S.L., whose partner and sole administrator is MAXIMILIANO MOBILITY INC. […] On February 1 2019, it acquired all the shares of PRESTIGE AND LIMOUSINE S.L. [the subcontractor, to whom the plaintiff was hired out via a temporary work agency].’
All of this illustrates that even courts, after having reviewed the parties’ arguments about the roles of various group members, may find it exceedingly difficult to pinpoint the company responsible for the employer functions which can essentially be ascribed to the platform's algorithms. It may thus be imagined just how difficult it can be for a claimant to correctly address a lawsuit.
Subcontractors
Regular subcontractors
Subcontracting has long been an established business practice in the platform economy, and notably in the ride-hailing sector, as some countries’ regulations would have made it impossible for Uber and similarly operating platforms to operate based on their original concept of working with self-employed individual drivers (e.g., due to regulations regarding licences for taxi services). 33 In other countries, that original model has become at least questionable, including with a view to the risk of reclassification by the courts. The latter is the case in France, where the Cassation Court has confirmed that working with self-employed platform drivers can be perfectly legal, 34 but also amount to bogus self-employment. 35 A growing body of case law evidences that ride-hailing platforms maintain a multitude of contracts with formally independent providers that manage (parts of) their fleet. Such subcontractors typically provide their drivers with vehicles and smartphones with the relevant app installed in a modified version, which ensures, notably, that the income generated is billed and paid to the subcontractor, rather than the individual driver.
French case law evidences that the difficulties and counterarguments faced by drivers wishing to be recognised as bogus self-employed and actually in an employment relationship with a platform's subcontractor are virtually indistinguishable from those experienced by those aiming to prove an employment relationship with the platform itself. A considerable number of decisions, including recent second-instance rulings, 36 have rejected such claims in relation to subcontractors of Uber and/or Kapten based on essentially the same argumentation used by the courts that have refused to recognise drivers as employees of Uber. In all of these cases, the drivers did not even attempt to argue that the platform company should be viewed as the actual employer. This includes cases in which the subcontractor was already declared insolvent at the time of litigation, 37 partly after pretending to plan to sign an employment contract with the plaintiff but never even proceeding to pay them for the work already performed. 38
In other cases, bogus self-employed plaintiffs have succeeded in being recognised as employees of a platform's subcontractor by the French courts. Just like Uber itself, its subcontractors have also at times been assessed differently even by (different chambers of) the same court. Accordingly, the Sucy en Brie Civil Court refused to requalify Uber drivers who had worked for the company EURL Car Park Service in 2019, while the Paris Labour Court confirmed those drivers’ employee status in 2020, and both decisions were upheld by the Paris Appeals Court in 2022 and 2023, respectively. 39 The judgments finding in favour of employee status deduced this from a combination of elements which logically follow from the drivers’ obligation to work via the app (such as the obligation to register as a self-employed driver, the prohibition of picking up other passengers and the deduction of Uber's commission from the payment) and others which were imposed by the subcontractor (such as the imposition of EUR 100 penalty for the cancellation of commitments, the deduction of the fee for using the car, the duty to wear a uniform, and the automatic extinction of the right to use the car when a driver is deregistered by Uber). A similar approach has been taken in other rulings that have confirmed the employer status of an Uber subcontractor 40 and effectively also in the Cassation Court's overturning of a decision in which it was ruled that a Kapten subcontractor's drivers were genuinely self-employed. 41
None of these decisions attached legal relevance to the fact that a driver's overall situation of subordination partly derives from instructions, supervision and sanctions imposed by the platform company. The same is true for the numerous proceedings in which a driver's status as an employee of a subcontractor 42 (or of a temporary work agency which hired them out to such a subcontractor) 43 was common ground between the parties. Most of these latter cases were taken in Spain, and virtually all of them featured questions of unfair dismissal and/or working time with a pronounced EU law dimension (both regarding the delimitation of work and rest time 44 and the employer's duty to record working time 45 – see infra section 3). While not directing claims against the platform directly, some of these lawsuits did try to rely on the platform's crucial role in shaping working conditions – notably in a case brought against six companies, whom the plaintiff wanted recognised as a group of employers, seeing as they were all Cabify subcontractors and shared a common registered office and even the same sole administrator. The judges (in first 46 as well as second 47 instance) disagreed, holding that the undisputable fact that the defendants constituted a group of companies with tight economic ties did not imply that they were a group of employers in the sense of a case law concept which leads to joint and several liability of all group members for the claims of each other's employees (see infra at 2.2.1).
Delivery platforms have so far been much less likely than ride-hailing platforms to subcontract their core activities, and case law on such constellations is still scarce. For now, Sweden is the only country where a Foodora subcontractor that presented itself as an umbrella company for self-employed deliverers was requalified by a labour court, which found that it had acted as a temporary work agency that employed the deliverers and hired them out to Foodora. 48 In the meantime, the Spanish platforms in particular are growing more aware that, with a view to the existing case law and ongoing administrative proceedings, 49 a requalification of their bogus self-employed workers seems almost unavoidable. For example, a second-instance case of 2022 50 shows that, as from the date of entry into force of the ‘Ley Rider’, 51 Uber Eats started to work with subcontractors which provide its formally (supposedly) self-employed riders with employment contracts. In a pending competition law case before a commercial court, 52 Uber is trying to force its competitor Glovo to stop working with bogus self-employed riders for good. This indicates that there may soon be a surge in legal disputes in relation to platform subcontractors in the delivery sector.
The minority of judgments which include an assessment of the respective roles of a platform and its subcontractor(s) and the consequences for allocating the label of employer will be discussed in more detail infra (subsection 2.2).
Clandestine subcontractors
Several decisions handed down by administrative courts in France earlier this year 53 identify platform workers as employees of persons who had no formal contractual relationship with the platform company, but effectively played a role akin to that of the subcontractors discussed in the preceding subsection. In these cases, an intermediary also built on the elements of direction and control that are inherent in working via a platform and supplemented them with the provision of work instruments as well as (possibly) instructions. More specifically, those involved had created an Uber Eats account in their own name, passed the login information (and also equipment, such as the backpack with Uber Eats logo) on to an individual without a work permit, and paid the latter for performing deliveries via the account (or at least promised to do so).
In all these cases, the arrangement had been uncovered in the course of a larger scale police operation in 2021, which led to fines being imposed by the immigration office and subsequently confirmed in court upon appeal. In all of them, the courts were satisfied that the workers’ testimonials indicated that they had committed to work in return for pay, with the owners of the accounts deducting a commission (e.g., 30%) from the income that Uber paid them. The judgments are generally very succinct in terms of assessing subordination, e.g., regarding whether the worker was in any form obliged to perform a minimum amount of work for the account holder, or rather just received their share whenever the account holder was paid by Uber.
In any event, these cases raise a number of questions as to how contractual relations would need to be assessed if Uber's relationship with its food delivery riders was generally requalified as an employment relationship (which has so far been rejected by the French courts 54 but confirmed e.g., by their Italian 55 and Swiss counterparts, 56 and French case law also includes countless examples of food delivery platforms being identified as employers). 57 In such a case, who would qualify as the platform's employee – the owner of the account or the person performing the work, both or none of them? Would it be possible for the account holder to simultaneously appear as an employee and an employer? These questions seem all the more relevant as clandestine account sharing is reportedly widespread, having also been at issue in a number of court decisions in other countries. 58
Temporary work agencies
Apart from subcontractors that were alleged or found to have in fact operated as a temporary work agency in the lawsuits (see also infra subsection 2.2), a number of Spanish cases 59 also involved players that already had the official status of a temporary work agency and hired drivers out to platform subcontractors. The involvement of such actors generally did not impact the outcome of the case significantly, as the claimants were entitled to equal treatment in the user undertaking (see Art. 5 of Directive 2008/104/EC) and the agency regularly employed them for precisely the period of time for which they were hired out to a specific company. One case 60 confirms that, in case of legal temporary agency work, the liability of a user under Spanish law is only subsidiary, not joint and several.
Clients
Another group of actors that may exercise employer functions in relation to a platform worker are clients that request a service via a platform. In this regard, it must be pointed out that platform models are, self-evidently, highly diverse. Notably where the platform's role is genuinely limited to that of a labour market intermediary, 61 an individual finding stable employment via its services will hardly be seen as a platform worker, and the client will be the only reasonable addressee of a claim of work in subordination. 62 There is, however, a broad grey area between these cases and those in which the platform seems the only reasonable addressee (notably where clients typically just order a single micro task via the platform and may not even know the identity of the worker(s) that completed it). The most difficult cases may be those in which a worker performs services that benefit a multitude of clients, but which are not pre-determined or standardised by the platform, so that the individual instructions of each client dominate the actual work performance.
From the scarce case law on these constellations, it seems clear that plaintiffs claiming employee status for platform workers would generally rather identify the platform as the employer. The decision-making bodies, however, have at times denied such claims and identified clients, who were generally not part of court proceedings, in absentia as employers or user undertakings in relation to platform workers. One example is cleaners, whose work – and remuneration – may be determined entirely between them and the client, with no or minimal interference by the platform. Accordingly, in a Norwegian case 63 the clients (private households) were identified as employers of cleaners working via the platform Vaskerhvitt, and a similar ruling was made in the first-instance decision 64 in a Dutch case on Helpling. The Dutch decision was modified on appeal though, 65 with the platform being classified as a temporary work agency and the households as user undertakings. Thereby, the judges acknowledged the structural role of the platform on which the cleaners relied in terms of the organisation and various guarantees, as well as the responsibility of the client for defining and controlling the individual service performed. This ruling is currently being challenged before the Dutch Supreme Court.
By contrast, ride-hailing and delivery platforms (i.e., the platform types addressed in the vast majority of platform work-related cases) generally offer a highly standardised service, in which clients appear to be more like passive recipients and are virtually never seriously considered as potential employers or users. The exception is a Belgian administrative decision 66 which identified the Belgian Platform Rider Association, of which clients (passengers) needed to be members in order to use Uber's app, as a co-employer, and thus liable jointly with Uber BV for the drivers’ social security contributions. This decision was later overturned in court, though. 67 The same happened in seemingly the only decision which construed a delivery platform (Uber Eats) as a temporary work agency and the restaurants as user undertakings. It was issued by a Swiss employment office, 68 confirmed on appeal, 69 but overturned in last instance by the Federal Court. 70 The last-instance decision stressed that any obligation of the riders to follow instructions by the restaurants concerned only minor elements of the service, and was based on the terms and conditions of the platform as the actor which actually established all relevant rules.
Liquidators and guarantee institutions
Finally, it should be mentioned that the party which ultimately appeared as the defendant in a good share of the cases on labour rights claims of platform workers was actually a liquidator of a company in a state of insolvency, or a wage guarantee institution in the sense of Art. 3 EID, whose responsibility for guaranteeing workers’ entitlements was equally dependent on those workers being qualified as employees. This situation is not unusual in cases where it has been claimed that a platform company itself was an employer (the most outstanding example being Take Eat Easy, which has been subject to over a hundred of lawsuits in France 71 long after it was declared insolvent); but even more prevalent where the purported employer was a subcontractor. Judgments assessing the employer status and/or obligations of an already insolvent platform subcontractor have been issued in Austria, 72 and notably in France 73 and Spain (where in one case 74 both the platform and its subcontractor were insolvent when the judgment was handed down).
What is striking about some of the cases is that there seems to be a fundamental lack of transmission of information between an insolvent platform employer and the liquidator or guarantee institution that has to argue against entitlements claimed by employees. For instance, platform workers have been able to claim outstanding remuneration simply based on a legal presumption of full-time work in cases where liquidators could not produce any useful evidence on the actual hours worked. 75
Focusing on the ‘platform vs. subcontractor’ issue
As indicated above, there have hitherto not been many lawsuits that forced the respective decision-making body to choose between a platform and its subcontractor(s) when allocating the role of employer and/or user undertaking in relation to a platform worker. In only a handful of them (a third) it was held that the interposition of a subcontractor must be disregarded, and that subordination existed directly between the platform and the worker.
Ride-hailing platforms
Claims of such a direct, disguised relationship of subordination vis-à-vis a platform have most frequently been brought in Spain, notably by employees of one of the numerous subcontractors of Cabify. These claims have regularly been rejected, including in cases where the subcontractor was fully owned by the platform company and had the same registered office. 76 In this respect, courts tend to stress that it remains perfectly legal for groups of companies to organise their business structure in such a way, and that the parent company provides its subsidiaries with a tool (a sophisticated digital platform) which enables the latter to optimise their staff management. This applies as long as there are no additional elements as developed in the national jurisprudence to identify a group of employers (with joint and several liability of all members for each other's employees) – including, notably, indications that there is no proper separation of each company’s financial assets, staff etc. 77
The plaintiff of one particularly noteworthy case 78 had identified a total of eight defendants, which included (in addition to the employer insolvency guarantee institution Fogasa and the Ministry of Finance) the platform company Cabify, its business partner Auro, and three companies that offer ride-hailing services via the Cabify app and belong to the Auro group, with two of them indicated by the court as fully owned by Cabify and Auro, respectively. The final defendant JT Hiring was the contractual employer (a temporary work agency named in other cases brought by Uber or Cabify drivers) 79 which had hired the plaintiff out successively to the three subcontractors. The first-instance court found a group of employers among three subcontractors, which had the same office, work centre and managing director and subjected their drivers to identical terms and conditions, whereby the plaintiff indicated a lack of clear separation between them by reference to a period in which he used a car owned by one of them in order to provide services for the account of another. This supported the lower instance court's conclusion that the dismissal of the plaintiff for economic reasons (outlined by the temporary work agency after the last subcontractor had terminated the hiring-out agreement) had been unlawful, seeing as one of the other subcontractors was hiring a number of drivers in the same period. The court of second instance, while agreeing that the dismissal was unfair, considered this the sole responsibility of the third subcontractor and found that there had not been a group of employers.
This illustrates the considerable hurdle which claimants need to overcome when attempting to bring claims against anyone but the contractual employer and/or, potentially, a formal user company – and all the more so when directing such claims against a platform company. It should be noted that in the case just described, Cabify and Auro had already been found to manifestly lack passive standing by the court of first instance. The judgment which gives the probably most detailed explanation in this regard is a Labour Court ruling of late 2022, 80 which overturned a penalty imposed by the Valencia labour inspectorate on Cabify and three of its subcontractors for illegal assignment of labour. The judge rejected the conclusions of the labour inspectorate, which argued that the business activity of the subcontractors had been entirely imposed by Cabify, and that their genuine entrepreneurial risks and opportunities had been severely limited due to the dominance of their activity by the platform rules, which imposed the pricing, took care of promotion and client support, and determined the instruction, control, rewarding and sanctioning of drivers. These arguments were reviewed in light of past higher-level case law and found to be insufficient to support a finding of illegal assignment of labour. According to the judge, the latter would require that the subcontractors’ role be entirely negligible – an assumption that was contradicted by the fact that the companies were responsible for obtaining the licences, for purchasing, maintaining and assigning the vehicles, for recruiting and dismissing workers, and for instructing them where and when to work. The judgment also stressed that the drivers’ geolocation via the platform was restricted to periods in which they were logged on, whereas the subcontractors could always geolocate their cars via telemetry.
This clear tendency of the Spanish judiciary can be contrasted to an Austrian decision that appears to be the only judgment which confirms the existence of an illegal assignment of workers to a ride-hailing platform by its subcontractors. More specifically, the pertinent decision of the Vienna Labour and Social Court, handed down early in 2020, 81 identified Uber as a user company in an illegal temporary agency work construct, resulting in joint and several liability for claims from drivers’ employment contracts with subcontractors whose insolvency had already been established. The decision was annulled on appeal, though. The Vienna Upper Regional Court, 82 while expressly considering the findings regarding the illegal assignment of work plausible, refrained from making a pronouncement on the issue, because in any event the court of lower instance had incorrectly identified a case of joint and several liability. 83 Despite this outcome, which meant that Uber formally won its appeal, the judgments evidence a considerably more generous approach than in Spain to the reclassification of contractual relationships, which in Austria can be established based on any one of four possible criteria that indicate that the attribution of employer status does not correspond to the ‘true economic content’ of the arrangement. The fact that those four criteria include, notably, the absence of an independent purpose of the formal employer's business and the worker's organisational integration in the other company's business have led Gruber-Risak 84 to conclude that ‘it cannot be disputed’ that Uber's relationship with its subcontractors amounted to illegal temporary agency work.
A similar line of argumentation is followed in an opinion of the Swiss Secretary of State in economic matters, which was requested by trade unions back in 2018. 85 It found that Uber was giving precise instructions to the drivers, effectively treating them as its own personnel, and should accordingly be seen as a user company that hired that drivers out from its subcontractors in an illegal temporary agency work arrangement. To this date, the opinion does not seem to have been followed up by the institutions, seeing as Uber subcontractors are still operating without a temporary agency licence without authorities intervening. 86
Finally, the apparently only published French case in which the plaintiff demanded Uber to be identified as co-employer next to its subcontractor 87 is noteworthy in that this part of the claim was rejected for the sole reason that the plaintiff had sued the ‘wrong Uber’. The judges shared the opinion (see supra at 2.1.1) according to which Uber France SAS’ role was essentially limited to assistance, support and marketing, whereas the configuration of the platform was the sole responsibility of Uber BV (which was not party to the proceedings). This underlines just how important that aspect of defendant identification can be, while at the same time preventing the judgment from being particularly instructive regarding the conditions under which French law might confirm a platform's responsibility for drivers hired via a subcontractor. Another striking example of a court rejecting a claim for subordination along the lines of ‘at least not with this defendant’ concerns a driver who had brought his claim exclusively against the subcontractor who had hired him for supposedly self-employed work. The judge found that (author’s translation) 88 ‘what [the claimant] considers to be a power of sanction and control was not directly exercised by the [subcontractor] but by the […] platforms Uber and Kapten, with which he has his own accounts.’ This formulation leaves open the question of whether the appeals judges basically considered the sanction and control powers sufficient to indicate an employment relationship, and if so, whether they would accordingly have identified Uber and Kapten as employers if only they had been included in the claimant's lawsuit. If the latter is not the case, the probably even more concerning conclusion would be that, by allocating different parts of what would in combination amount to an employer's function to a platform and its subcontractor(s), respectively, the requalification of a worker subject to such a ‘fractioned’ form of subordination could be successfully prevented.
Other platforms
The case law on delivery companies also confirms the strict approach taken by the Spanish courts. In particular, the platform Glovo – which has been identified as an employer of its directly hired workers in over a hundred of lawsuits 89 – was absolved of any responsibility for the employment-related rights of a rider hired as a bogus-self-employed by a subcontractor. 90 Ditto for Amazon, whose employer status in respect of directly hired delivery drivers has recently 91 been confirmed in court for the first time, but rejected in a case 92 where the platform had subcontracted deliveries to a company that hired its drivers. Again, the courts in first and second instance acknowledged the unmistakable dependence of the subcontractor on the platform company, for which it worked exclusively – to the degree that once Amazon terminated its contract with the subcontractor, the latter ceased its activity and dismissed, inter alia, the plaintiff of the case at issue. Yet, once again, the judges stressed that the subcontractor did exercise a relevant business activity that could be separated from Amazon's. This came down to purchasing, maintaining and assigning the vehicles (which featured Amazon's logo), hiring, firing and instructing the drivers (who wore Amazon uniforms) and choosing among the ‘routes’ (groups of orders with similar destinations, which Amazon's algorithm deemed fit to be completed by one driver with a vehicle of a specified cargo capacity within a specified timeframe). The fact that, once a ‘route’ was assigned, the driver's activity was entirely steered and controlled by the platform and, if needed, Amazon's support centre, was found insufficient to make the subcontractor's role appear irrelevant. Consequently, the judges found no illegal assignment of work, and no responsibility of Amazon for the claims of the driver (which included back pay of wages in accordance with the applicable collective agreement and compensation for unfair dismissal).
Again, courts in other countries have arrived at different conclusions, albeit under very particular circumstances. More specifically, the delivery platform Stuart had hired some of its riders via a subcontractor that was so devoid of any business activity that the Paris Criminal Court found its ‘existence purely theoretical’. 93 In Italy, the Turin Civil Court was confronted with the perhaps most disturbing example of abusive hiring-in practices by a platform that has reached the courts so far. The subcontractor in question was a sole proprietorship, run by an individual who, with full knowledge and endorsement of Uber Eats management, did exactly what platforms famously avoid doing, namely, coercing riders into working specific shifts at peak times, by means that included threats, insults, and the demonstrative immediate removal of riders who talked back to him. This treatment, as well as the fact that the riders – mostly migrants, some of whom did not have work or residence permits – were severely underpaid, led to a condemnation for exploitation of labour by the Court of Milan. 94 Regarding the contractual relationship, the detailed account of the facts by the Court of Turin 95 evidences that the subcontractor exercised genuine authority over the workers, including, notably, hiring and firing as well as issuing (clear and aggressively formulated) instructions and imposition of sanctions. Yet, the court found, based on the records of communication between Uber and the subcontractor, that all of the actions were effectively just implementing instructions that the subcontractor had received from Uber, including very specific reprimands about the failure to ensure a sufficient number of riders on a particular occasion and occasionally even ‘recommendations’ on how to make riders adjust their working hours to meet Uber's needs (suspension of payment). In combination with the fact that the subcontractor clearly did not have the possibility to supervise, let alone intervene in the riders’ activity on the platform (instead being informed by Uber if the performance of ‘his’ riders was unsatisfactory), the court concluded that the interposition was artificial and that actual authority was exercised by Uber itself. Lastly, in Sweden, a claim brought by a union successfully achieved a reclassification of a Foodora rider as an employee. The court, however, disagreed that the platform should be seen as a direct employer: rather, Foodora was acting as a user undertaking, and the contractual employer was the subcontractor that had managed the (bogus self-employed) riders’ pay, tax and contributions. 96
Apparently the only case outside the ride-hailing and delivery business (which has always dominated the case law on status questions in platform work) 97 concerned the Spanish platform Joyners, dedicated to assigning personal care workers. While the core of the judgments assessing the platform in first and second instance 98 focused on the identification of bogus self-employment (which was confirmed), the rulings also examined the role of a subcontractor. 99 This role had evolved over the years of the platform's operation, as the subcontractor had originally admitted care workers as working members of a cooperative until 2017, when the platform company made all care workers sign a standard cooperation agreement for supposedly self-employed services, which reduced the subcontractor's role to one of providing billing and invoicing services for those workers that joined the platform via its intermediation. It was only from this point in time that the judges recognised the subcontractor's role as being negligible, resulting in a requalification of all care workers as employees of the platform from 2017 onwards.
Main considerations
The decisions which opt to choose a platform company over its subcontractor(s) when allocating the status of employer 100 basically need to ascertain two facts: first, the existence of clear elements of exercise of employer functions by the platform; and second, that the subcontractor's role is comparatively negligible or in any event less decisive in characterising the situation as one of subordinate work.
The first of these elements requires an assessment similar to that performed in the much greater number of cases concerning only the bilateral relationship between a platform and its employees. It should be noted that with the sole exception of Cabify, all of the platforms at issue in the cases involving a subcontractor also work with directly contracted individuals, and have already been reclassified as an employer of those workers in at least one country.
101
Accordingly, this part of the judgments uses arguments entirely comparable to those advanced in the ‘bilateral cases’, with an emphasis on the platform's meticulous organisation and control of the work performed. This includes
- the processing of individual client requests (for a ride, delivery, or personal care provision) to enable optimal matching with a worker ready to provide such services; - the determination of the price and processing of the payment; - geolocation and real-time tracking of task completion; - management of support, advice, troubleshooting, complaint mechanisms and dispute resolution regarding virtually all aspects of work performance; - control, supervision and documenting of performance.
Additionally, the rulings point to the various forms in which platforms are able to steer without giving actual commands – by incentives, gamification, and straight-out sanctioning regimes. This includes, notably, mechanisms that instrumentalise client expectations, as deviations from what is advertised by the platform are likely to be reflected in bad ratings or even complaints, which the platform company in turn uses to apply measures e.g., in respect of remuneration, access to reservation systems or even the platform as such. A particularly prominent feature is also the conditional granting of advantages to workers who meet certain performance criteria, which appears more like a sanction for everyone else when failing to obtain them actually makes working through the platform unviable as a major source of income. This is complemented by the less subtle, straightforward sanctions for behaviour that is expressly discouraged or prohibited by the platform, but this never concerns issues that have traditionally been considered the core of an employer's prerogative – such as where, when and how much to work.
In this regard, the most interesting case is probably that of the Joyners platform. Other than the ride-hailing and delivery businesses at issue in the other cases, the nature of personal care services does not allow for the offer of a highly standardised service in which all essential details are pre-determined by the platform. Instead, these details can only be properly determined in the relationship between the carer and the care-dependent individual, based on the client's needs and the carer's expertise. In casu, there was also no non-compete clause or duty to wear the company's logo. Yet, as the courts pointed out in first and second instance, 102 the essential contents and pricing of any specific service were pre-determined by the platform, with the worker only able to accept or refuse it in its entirety. Joyners also had detailed requirements regarding carers’ qualifications and experience, monitored service provision, and presented itself as the guarantor of the quality of the services vis-à-vis clients. It also processed the payment and offered support and dispute resolution services. Eventually, the Spanish courts came up with a ruling resembling that of the Dutch court of appeal judgment in Helpling, 103 namely, that there had been an (illegal) assignment of work, with the platform acting as a temporary work agency.
In a second step, the domination of work performance by the platform company needs to be juxtaposed with the role played by a subcontractor who appears as the worker's direct contract partner. This role may appear particularly limited in cases where workers are performing their activities in a relatively autonomous way, similar to those who contract with the platform as individuals. By contrast, a subcontractor's role becomes much more employer-like as soon as they add rules to those established by the platform. This is virtually unavoidable in those constellations where the subcontractor provides tools and equipment (notably vehicles), which requires rules on aspects such as use, maintenance etc. Much more crucially, a subcontractor can add the one element which has emerged as the greatest stumbling block for gig workers seeking to qualify as employees: 104 an obligation to work. In this respect, the arrangements brought before courts in various countries exhibit a considerable variation, ranging from clearly defined obligations regarding time, place and number of tasks to be performed through the platform (expressed e.g., in terms of task refusal and cancellation rates) to more generic frameworks or targets.
For instance, from the French and, notably, Spanish case law on ride-hailing platform subcontractors, it appears that the typical arrangement is for the worker to be assigned a vehicle, which they may or may not be allowed to take home and use for private purposes when reimbursing fuel and associated costs. Arrangements regarding the aggregate weekly working time (possibly within a reference period) are common; as are instructions to stay within a certain geographical area (e.g., within the boundaries of M-30 orbital motorway in Madrid). By contrast, fixed work hours are imposed only exceptionally, and strict obligations to accept all tasks as offered via the platform are almost never invoked. Regarding pay, both fixed monthly rates and performance-related pay, as well as combinations of both, are described.
If both the platform and the subcontractor are found to fulfil some but not all criteria indicative of employer status, the outcome can range from a finding of some form of shared liability of both parties (based on legal institutions such as co-employment, undeclared temporary agency work or subcontractor liability), via the choice of one entity as the only liable party, to a denial of reclassification. Needless to say, the latter scenario seems particularly concerning when the worker is effectively in a situation of subordination which just cannot be attributed to a single source of authority. At least when taken together, the judgments evidence that in some cases it can be argued that role of both the platform and the subcontractor lacks key elements of classic employer status. Taking the example of clandestine subcontracting as described supra at 2.1.2.2, it is noteworthy that the courts in those cases rejected the arguments of the defence essentially because, based on the testimonials from police records, they did not believe that the workers were not obliged to work. Yet, irrespective of what actually happened in the cases subject to those proceedings, it seems realistic that in many cases of clandestine subcontracting the subcontractor simply shares their account data and pays the worker a share of whatever they are paid by the platform. In such a scenario, the basis for qualifying the account holder as an employer would arguably seem a rather weak. At the same time, the platform – which exercises all of the instructing, supervising and controlling functions that the account holder lacks – does not choose or even know about the identity of the worker, nor determine their wage. The situation actually resembles one of temporary agency work, whereby it seems unclear whether national laws on temporary agency work could apply to a situation in which the user undertaking is not even aware of the assignment of labour.
‘Wrong employer’: does it matter? The claimed entitlements at issue
The above discussion of the complexities of identifying the entity or entities having exercised actual employer functions in multiparty arrangements prompts the question of whether it actually matters. At first sight, this is obvious only for cases involving a penal or criminal sanction, as a sanction obviously fails in its purpose when imposed on the wrong subject. In this regard, the Italian Uber Eats case 105 is exceptional, in that the courts had access to a comprehensive transcript of communications between the platform and its subcontractor which enabled it to identify Uber both as the employer and as co-perpetrator of the exploitation of workers. For lack of such evidence, it seems clear based on the example of this case alone that using subcontractors can provide an avenue for platforms to escape liability for malpractices that are a clear result of the subcontractor being put under pressure to meet the platform's demands.
What appears less obvious is whether, for a platform worker who has been recognised as an employee in court, it makes any difference who has been identified as their employer. In many of the cases discussed in this contribution, this is not entirely clear. Many claimants were found to be entitled to the payment of wages and other financial entitlements, often including compensation for unlawful dismissal, with the amount established based on the law or a sectoral collective agreement, and no indication that those entitlements would have differed if another entity had been identified as the employer.
There are, however, a number of indications that the identification of the employer can be highly relevant also from the worker's perspective.
Probably the most self-evident aspect is that of insolvency. As indicated supra at 2.1.5, cases of bankruptcy are by no means isolated incidents in the labour platform economy. And while the risk may also be non-negligible for ‘direct employees’ of a platform (as evidenced by Take Eat Easy or Helpling, and also the oft-cited 106 fact that even Uber turned a – modest – profit for the first time in its existence only in the second quarter of 2023), it seems clear that the probability of insolvency is greater if the employer is an SME that is in most cases entirely dependent on one client (the platform) (note the prevalence of sole proprietorships among the defendants in the cases described above). The consequences of a subcontractor's insolvency depend on national law. As can be seen from the cases discussed, in Spain a regulation is in place which regularly ensures the main contractor's liability. This led one Spanish second-instance court to overrule the court of lower instance's finding of illegal temporary agency work, while stressing that this change was in fact irrelevant for the outcome, as Cabify was in any event jointly and severally liable for the worker's rights. 107 The exact opposite occurred in Austria, where an appeals court indicated that the finding of illegal temporary agency work was most likely correct but did not need to be examined further, as even such an illegal assignment would make Uber at most subsidiarily (and not jointly and severally) liable. 108
In countries where working conditions are crucially dependent on the applicable sectoral collective agreements, it is the categorisation of the employer which determines the minimum standards applicable to all staff. This issue has arisen in a number of Spanish cases, 109 as well as in the Austrian case mentioned in the previous paragraph, in which the plaintiff sought entitlements in line with the sectoral agreement applicable to the temporary agency work industry. The fact that the driver expected significant benefits from the application of the agreement for that sector – which is certainly not among the most generous in terms of wage rates and benefits – indicates the level of income which these (employed!) drivers are generally generating through their activities. Ditto for the monthly pay rates expressly indicated in many of the judgments on Spanish private hire drivers – and this notwithstanding the fact that all of them were formally covered by a collective agreement. Perhaps even more importantly, many agreements (including the one on private hire drivers in Madrid, which was applicable to the greatest number of relevant cases) allow for modification at company level. 110 Here again, there might be a considerable difference between the bargaining power of workers’ representatives, depending on whether the company at issue is a large and prosperous platform provider or an SME forced to work with tight profit margins due to its own weak bargaining position vis-à-vis the platform on which it depends economically.
Company-level bargaining is another area where legal rules are likely to differ depending on the size of the company – resulting in a lower level of protection for the staff of SMEs. The German case law on platform workers features a number of cases regarding the right to elect a works council 111 – for which the applicable rules are crucially dependent on the platform undertaking being organised as one company rather than a network of SMEs among which the staff is distributed.
Another potential issue concerns seniority, notably in a context where it seems to be common practice - at least among the Cabify subcontractors party to the biggest group of cases - to have a particular driver work for a number of different companies in succession. 112 Several judgments indicate that in such cases the driver's seniority with another company of the group was fully credited, but it seems unclear whether different subcontractors of a platform are necessarily legally required to do this.
The issue of the lawfulness of dismissal has been raised in a large number of the cases. Here, economic dismissals in particular - at issue in several cases 113 - are self-evidently connected to circumstances at the employer, which in many cases differ significantly between the platform and its subcontractors. This is illustrated notably by cases in which courts are asked to consider the discrepancies in staffing practices of different subcontractors of the same platform (one hiring while the other is firing), 114 and e.g., also those in which a subcontractor's staff are dismissed as a result of the platform's decision to discontinue cooperation with this specific subcontractor, 115 at a time when the platform's business is basically expanding. Even more obviously, the case mentioned supra in the introduction 116 illustrates that splitting up a platform's workforce among a number of subcontractors can prevent the application of the rules on collective redundancies.
Finally, the only issue that is equally prevalent as dismissal compensation in the case law is that of working time. In that regard, what is striking in relation to the Spanish cases on ride-hailing companies is an approach eventually confirmed by the Spanish Supreme Court, 117 according to which workers appear essentially powerless to prove that they have performed unpaid overtime, night work and weekend work. Usually, plaintiffs have attempted to substantiate such claims by providing platform records of the periods in which they have been logged on – which, however, does not necessarily indicate readiness to work, provided the workers remained free to accept or reject individual ride offers as received via the platform. Where the employer has produced working time records in some form, the circumstances have not been found to be comparable to those in the CJEU judgment in CCOO. 118 Some judgments 119 clearly indicate that such records are presumed to be accurate even if they seem to be exclusively based on an agreement between the parties that the worker would perform exactly eight daily hours of work and no overtime. This begs the question of whether, in the hypothetical situation where the platform itself has been identified as an employer, it can rely on the same argumentation, given that, in view of typically very comprehensive collection and storage of rider data by platforms, 120 such an employer would be perfectly capable of providing evidence of the actual working time.
Subcontracting liability in a directive on platform work?
Subcontracting liability may be considered a powerful fallback solution to increase workers’ protection in multiparty constellations, in that it is not dependent on complex questions of the actual legal qualification of the contractual relations at issue. While it is by no means capable of addressing all of the issues discussed in the last section as consequences of the identification of an employer which is unfavourable to the worker (specifically: a subcontractor rather than the platform itself), it does at least offer a remedy for the perhaps most considerable risk of being unable to enforce acquired entitlements. This section will accordingly be dedicated to a description and brief discussion of minimum European standards of subcontractor liability, both de lege lata and – specifically with a view to subcontracting in the platform economy – de lege ferenda.
Precedents of EU-level regulation of subcontracting
At present, EU law contains specific provisions on subcontractor liability only in respect of cross-border situations.
For one thing, Art. 8 of the Employer Sanctions Directive (2009/52/EC) provides that ‘[w]here the employer is a subcontractor […], Member States shall ensure that the contractor of which the employer is a direct subcontractor may, in addition to or in place of the employer, be liable to pay: (a) any financial sanction imposed under Article 5 [i.e., for illegal employment of a third-country national (TCN)]; and (b) any back payments [of wages, taxes and social security contributions and repatriation costs].’ The Article allows for stricter subcontractor liability under national law, but requires also that ‘[a] contractor that has undertaken due diligence obligations as defined by national law shall not be liable under paragraphs 1 and 2.’
In a similar vein, Art. 12 of the Posting of Workers Enforcement Directive 2014/67/EU provides, with respect to posting in the construction sector, that ‘Member States shall provide for measures ensuring that in subcontracting chains, posted workers can hold the contractor of which the employer is a direct subcontractor liable, in addition to or in place of the employer, for the respect of the posted workers’ rights [to any outstanding net remuneration corresponding to the minimum rates of pay and/or contributions due to common funds or institutions of social partners]’, but ‘may […] provide that a contractor that has undertaken due diligence obligations as defined by national law shall not be liable’.
In other words, the mandatory European minimum standard is in both cases limited to the subcontractor's direct contractual partner. This appears to be a significant limitation, as it indicates that a main contractor wanting to escape liability just needs to ensure that its subcontracting chain consists of more than one layer, since – for lack of more stringent regulation at national level – they cannot be held liable for the entitlements of a sub-subcontractor's workforce. And while this limitation might be understandable in relation to subcontracting in the construction sector, where a main contractor might find it difficult to follow up on the staffing policy of all of a potentially very large number of sub-subcontractors involved in a project, it would appear effectively unjustifiable in relation to a digital labour platform, which has a clear and detailed overview of all workers connecting to it at all times, irrespective of the number of subcontracting layers involved.
A second element contained in both cited provisions concerns the possibility to avoid liability by complying with ‘due diligence obligations’ – mandatory in relation to TCNs, optional in relation to posted workers. Whether such a limitation of liability seems justifiable might depend on the specific actions required from a company to consider it to have acted with due diligence. For instance, if the main contractor requires regular proof of the subcontractor's actual monthly payment of wages and social security contributions, checks whether the amounts are in line with both the individual employment contract and the provisions of applicable laws and/or collective agreements, and takes action as soon as it appears that there are arrears in payment, it might seem fair to absolve it from liability for any irregularity which it could not reasonably have detected. At present, though, most Member States provide for an overall (ex-post) assessment of due diligence rather than clear criteria, 121 which makes it difficult to assess whether national law might offer too easy a way out of liability.
A noteworthy difference between the two cited provisions concerns their scope – with subcontracting liability for posted workers being limited to minimum wages and the corresponding social security contributions, while the full back payment is at issue in case of illegally employed TCNs.
A different approach is taken in the proposed corporate sustainability due diligence directive (CSDDD), 122 which is to cover all subcontractors in a company's value chain, but only for those claims for damages for adverse human rights impacts in the sense of Art. 3(c) of the proposed directive.
Finally, it is significant that the Commission's implementation report on Directive 2014/67/EU 123 notes that ‘Member States may not set up a system of subcontracting liability which applies only to cross-border service providers, as this would directly discriminate against those service providers. Such system could might [sic] discourage customers’ final clients and contractors to contract or subcontract with undertakings established in other Member States.’ In other words, Member States are actually already considered obliged to create a general scheme for subcontractor liability, which applies also to purely national situations involving no posted workers. And while this could in principle be limited to the construction sector, it appears that the law of most Member States does not envisage a sectoral differentiation. 124
A first look at the European Parliament's proposal
The European Parliament is entering trilogue negotiations on the proposed directive on working conditions in platform work with a draft 125 that deviates from the Commission's proposal 126 in a number of respects. Without there being a need to dwell on other elements (many of which are discussed in other contributions of this special issue), a particularly intriguing suggestion is that which seeks to insert Article 12b – ‘Subcontracting liability’. It stipulates that ‘Member States shall, after consulting the relevant social partners in accordance with national law and practice, provide for measures to ensure that in subcontracting chains persons performing platform work have an effective remedy whereby the digital labour platform of which the employer is a subcontractor can be held liable, in addition to or in the place of the employer, for any infringement of the platform workers’ rights provided for in this Directive, including with respect to any outstanding remuneration and contributions due to the common funds or institutions of the social partners’ (emphasis added).
This formulation appears rather confusing, notably as the proposed directive does not provide for rights to remuneration or contributions. The rights which could be originally based on the directive – such as those relating to algorithmic management or to information and consultation – are not necessarily monetary in nature. If the idea is to oblige platforms to provide, e.g., information on their algorithms directly to workers hired via a subcontractor (and their representatives) – which might seem a very reasonable proposal – then the term ‘liability’ is perhaps not ideal for clarifying this. If, in turn, the provision means to imply that a platform should be liable for all workers’ rights (which seems more likely), including all entitlements in the employment contract agreed between the subcontractor and the worker (irrespective of whether the platform could reasonably have known about them by any standards of due diligence), it would seem highly recommendable that this be clarified the formulation.
What seems very clear, though, is that (unlike the provisions described in the previous subsection) the platform's liability is meant to apply not only to staff of a platform's direct subcontractor, but to all ‘layers’ of subcontracting. As illustrated by the complex contractual relationships at issue in the cases explored in this article, this seems a very necessary element to ensure the effectiveness of the provision. The other key deviation consists in the absence of any reference to a due diligence defence.
As regards the latter aspect, the question of whether or not some form of due diligence defence should be included in the final version might also depend on whether the desired outcome is to discourage subcontracting, or whether the idea is that subcontracting can in principle even be beneficial for workers (see next subsection). In the latter case, one might consider giving the platform a ‘safe option’ to protect itself from liability by exercising some form of stringent control (such as demanding rigorous evidence of regular monthly wage and contributions payment). If the due diligence standard is clearly defined, it seems conceivable that this could even result in a greater incentive for the platform to design its relationships with subcontractors in a way that genuinely minimises the risk of the subcontractor defaulting while owing significant amounts to its workers.
Conclusions
As shown by the mounting relevance of multiparty relationships in the case law explored in this article, labour relations in the platform economy are regularly more complex than might be expected based on the purely bilateral questions raised in many academic and policy discussions. Most notably, as mentioned, there are indications that subcontracting is on the rise, inter alia, as a result of the awareness of platforms of the mounting ‘risk’ of being requalified as an employer themselves.
When discussing how to approach the multiple problems and hurdles that multiparty relationships can entail for the enforcement of labour rights, it seems important to stress that the existence of subcontractors is not per se bad news for working conditions in the sector. Platforms are famous for offering the workers they contract with directly a considerable degree of freedom in terms of when and how much they work, and what tools to use. As has been discussed extensively, 127 many workers appreciate these freedoms; for some, it may even be the only way of work that is compatible with their current circumstances for various reasons (work, studies, family obligations, health conditions etc.). However, as indicated by the hundreds of lawsuits brought by platform workers claiming employee status, many of them would gladly trade those freedoms for a stable employment relationship with a defined monthly income, predictable working hours and equipment that is provided and maintained by the employer. This is precisely what subcontractors in many (though by far not all) of the cases are offering. Seeing as the cases frequently feature claimants with obviously limited financial means or debts, it seems clear that many would not even be able to purchase and maintain the necessary equipment (e.g., a car that meets the standards of a ride-hailing or delivery platform) by their own means.
What is more, subcontractors may be instrumental in a certain diversification and ‘democratisation’ of a market that is generally dominated by a very small number of large players. Notably, the cases show that at least some subcontractors work with different platforms, 128 which might prevent their workers from losing their jobs and the ratings they have built up with a platform if it goes bankrupt (as e.g., Take Eat Easy and Helpling) or leaves a particular national market (see e.g., Deliveroo's departure from Spain and the Netherlands or Uber's recently announced 129 departure from Italy). In such a case, even a subcontractor that has so far worked exclusively with the platform concerned would likely be able to secure much better conditions for its workers when switching to a different platform, as competing platforms may be highly interested in taking over the assembled fleet the subcontractor has at its disposal and thus willing to provide, e.g., recognition of the workers’ experience with the previous platform. It should be noted that in Spain – where the majority of ‘subcontractor cases’ have been heard – it appears that fleet managers such as Auro have gained considerable market power, which makes Cabify, Uber and Bolt compete for the option to contract with them. 130
At the same time, the case law discussed in this article illustrates the multitude of legal issues that come with the increasing prevalence of partly very complex multiparty relationships in platform work, both in terms of the difficulties around identifying ‘the employer’ when workers appear to be working in a situation of subordination and in terms of the enforcement of workers’ rights.
As regards the former, it seems noteworthy that laws in several countries apparently enable decision-making bodies to declare a company liable for workers’ entitlements as an employer without needing to establish that the company at issue has exercised the role of employer in respect of all aspects of the employment contract. For instance, a Belgian administrative decision (which was later overturned in court) 131 described different elements of subordination that related partly to Uber BV and partly to the Belgian Platform Rider Association, before concluding that both should be seen as employers for purposes of social security. A Norwegian decision 132 in turn found, in relation, to a cleaning platform that the clients (private households) needed to be seen as employers as regards Occupational health and safety obligations, while stressing that this did not necessarily preclude the identification of the platform as the employer in relation to other duties. Yet, findings of this kind have so far been isolated, and the overall impression is certainly that the complexities of the relationships between the parties increase the procedural obstacles which plaintiffs are confronted with. Thereby, failing to identify the right company as a defendant will in the best case mean that the claim needs to be re-submitted against the ‘right’ company, resulting in additional costs and waiting time; and in the worst case, the claimant might miss the window in which they must bring claims against the correct defendant. Including several defendants ‘just in case’ is in turn likely to increase the costs and duration of the original claim. Most concerningly, as described, there are at least indications that sufficiently ‘dispersing’ various employer functions among different actors can make it impossible identify any of them as the employer, even though the overall situation would indicate that work is performed in subordination.
As regards, the enforcement of labour rights, the analysis of the issues that have come up in case law so far indicates that several issues may require particular attention from a regulatory perspective, in order to avoid rights becoming unenforceable in a multiparty relationship.
For instance, rather than focusing only or mainly on in-house collective agreements signed with a specific platform, sectoral-level agreements may have a key role to play when a platforms’ workforce (in the broad sense) is spread across a number of subcontractors, who should ideally also be represented in the negotiations on adequate minimum standards for the sector. In addition, it may be worth thinking about a functional equivalent to company-level workers’ involvement with the platform in a situation where a major part or even all of its workforce is outsourced to subcontractors. In this respect, at least when the platform is a multinational and its subcontractors can be seen as part of a group of undertakings, the workers are basically entitled to initiate the establishment of a European works council at group level in accordance with Directive 2009/38/EC (EWCD). It is, however, unclear (and might merit further comparative investigation) if such potential group-level representation is actually equivalent to the forms of workers’ involvement that would be possible if those workers were part of one (platform) undertaking.
Other than the EWCD, the Collective Redundancies Directive 98/59/EC applies to ‘employers’. Accordingly, dismissals carried out by several members of a group of undertakings cannot be counted together when assessing the existence of a collective redundancy for purposes of the Directive. This applies even when such dismissals happen at the same establishment (note, e.g., Cabify's subcontractors tend to share even the work centres where their cars are parked). The same limitations may apply in respect of additional duties of the employer which national law imposes on the employer in case of collective redundancies (relating e.g., to collective bargaining, outplacement, and/or compensation). In this sense, it may seem worthwhile de lege ferenda – not only with respect to platforms – to consider the specific context of dismissals in a group of undertakings. This may include not only the criteria for establishing collective redundancies, but also the assessment of whether a dismissal is based on a just cause – which is in some way required by the national law of all European countries 133 – notably in situations where employers invoke economic reasons for a dismissal that coincides with the hiring of workers for the exact same work by other group members.
Finally, it appears somewhat absurd that in a context where platforms are collecting excessive amounts of data on their workers, 134 crucial information concerning working time in partiuclar is not available in proceedings to which the platform itself is not a party. This concerns disputes between workers and regular subcontractors as well as those involving any other of the possible defendants indicated supra in subsection 2.1, including liquidators and guarantee institutions. This lack of information can be beneficial for workers, particularly if it leads to their working time being presumed to have been full-time when it was actually less, as indicated in some cases (cf. supra at 2.1.5). However, as shown in a much greater number of cases, it works to the detriment of those workers who have worked overtime and could possibly claim additional rights for working at night, on weekends and holidays etc. This appears particularly alarming in light of other cases, which evidence that platform subcontractors are meticulously evaluating the performance/productivity of their workers based on the abundance of data collected from the platforms 135 (and might even use the services of another specialised subcontractor for this purpose) 136 and base dismissal decisions on the results. It goes without saying that this creates an incentive (or pressure) to work overtime, particularly if, as is virtually always the case in the ride-hailing business, the platform subcontractor makes vehicles and equipment available to its workers for (much) longer periods than the agreed working hours and gives them the ‘freedom’ to choose their actual working time within those periods.
In this sense, from a regulatory point of view, it would seem desirable to clarify (and not only for platforms) how the principles of the GDPR (including that of data minimisation expressed in Art. 5(1)(c) GDPR) should be aligned with the duty to record working time flowing from the CCOO judgment (see supra section 3) in a context where ‘availability for work’ is difficult to establish. Notably, when a driver or rider can technically be logged on to the platform without actually planning to accept ride or delivery offers, there would need to be criteria (which might ideally be established be collective agreement) to determine when the worker can legitimately be considered not working (e.g., in respect of the passing of time, non-acceptance or cancellation of work offers etc.). 137 Platforms might then be required to record precisely the information necessary to determine these elements of the work history, even when they are not the direct employer, as this information can be crucial for third parties – such as the workers, and also guarantee institutions - which should not be forced to pay wage arrears for full-time work to persons whose activity on the platform has in fact been sporadic.
In other words, various elements may need to be considered in a regulatory approach to multiparty relationships in the platform economy. At the same time, undeniably, subcontractor liability regulation might be the one key ingredient to prevent workers (and/or the publicly funded guarantee institutions) from having to foot the bill for the platform's work with unviable subcontractors, which are used as long as they are profitable but shield the platform from responsibility for workers’ rights’ violations. Accordingly, the European Parliament's initiative to include a clause on subcontracting liability in the proposed directive on platform work may be welcomed as a promising approach to tackling one of the core risks of subcontracting. As described supra in subsection 4.2, a clarification of the draft formulation would seem highly desirable, and the precise ramifications of liability may need to be assessed against the intended effects (as exemplified by the question of whether to include a due diligence defence, and if so, how it should be designed). In any event, the issue definitely seems deserving of particular attention the upcoming trilogue deliberations at EU level.
Footnotes
Declaration of conflicting interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
