1.1. Contract labour

The Employment Relationship Recommendation, 2006 (No. 198) has arguably been a long time in the making. The International Labour Office’s law and practice report for 1997’s Committee on Contract Labour, which was discussed at the eighty-fifth session of the International Labour Conference in 1997, initially explored the possibility to adopt an instrument on contract labour (ILO, 1996). According to the draft instrument’s definitions, the contract worker personally performs work under actual conditions of “dependency on” or “subordination to” the user enterprise (ILO, 1998a). Because subordination is commonly associated with a regular employment relationship, the novelty of the instrument, therefore, would have subsisted in the concept of “actual conditions of dependency”. Indeed, the draft instrument on contract labour did not initially address situations of “disguised employment”3 as such. Employment protections would, instead, have been extended to certain categories of dependent self-employed workers. However, the discussions on the definition of ‘contract labour’ never achieved conceptual clarity, which, in turn, proved to be a fundamental obstacle to the adoption of the instrument (E. Marín, 2006, pp. 340-341).4 Already in 1998, the Employers, in fact, proposed to invest time and resources in “a Recommendation providing guidelines and setting out procedures for dealing with disguised employment” instead.5 The concept of contract labour was abandoned, and the discussions were reoriented towards a less far-reaching objective, i.e. how to determine when a worker is in an employment relationship. The question was no longer which self-employed workers fare in conditions similar to an employee’s and are, therefore, “contract workers”. The question instead pivoted around what the characteristics of a “genuine employee” are. This more limited focus of Recommendation No. 198 has been criticised by some observers (A. Hyde, 2012, p. 98).

In 2000, a Meeting of Experts on Workers in Situations Needing Protection was convened to reassess the situation. It agreed “that the ILO can play a major role in assisting countries to develop policies to ensure that laws regulating the employment relationship cover workers needing protection.”6 This radically shifted the focus from the scope of ‘contract labour relationships’ to the scope of ‘employment relationships’. Whereas the instrument on contract labour was not supposed to deal with disguised employment, as such, this phenomenon became central for the prospective instrument on employment relationships. Furthermore, the ILO commissioned 39 national studies as a follow-up to 1998’s ‘Resolution concerning the possible adoption of international instruments for the protection of workers in the situations identified by the Committee on Contract Labour’.7 These studies and the conclusions reached by the Meeting of Experts of 2000 informed a report on the employment relationship's scope, issued by the ILO itself (ILO, 2003). Following a general discussion on the employment relationship's scope during the ninety-first session of the International Labour Conference in 2003, the ILO Governing Body was invited to consider future action on the employment relationship.8 In its 289^th session, in 2004, the ILO Governing Body decided to pursue a single ILC discussion on the “the employment relationship” with an aim to set an international standard.9 This process would eventually result in the Employment Relationship Recommendation, 2006 (No. 198).

1.2. Recommendation No. 198

Recommendation No. 198 was arguably adopted with one eye on the past and the other on the future. On the one hand, the Preamble refers to the traditional idea that labour law “seeks, among other things, to address what can be an unequal bargaining position between parties to an employment relationship”.10 On the other hand, the same Preamble recognises that international guidance on the employment relationship's scope should remain relevant over time, and that “protection should be accessible to all, particularly vulnerable workers”. The Preamble also emphasizes, for example, that it is essential to establish who is an employee and what rights this employee has in relation to a given employer in the framework of transnational provision of services. This balance between labour law’s traditional rationales and contemporary challenges sees itself reflected throughout the instrument.

The instrument’s primary purpose is to have the ILO Member States formulate and apply a “national policy for reviewing at appropriate intervals and, if necessary, clarifying and adapting the scope of relevant laws and regulations”. It thereby adopts a “processual approach”, in which “an ongoing process of regulatory evolution” takes place (M. Freedland, 2007, pp. 18-19). As part of this “long-term” national policy (E. Marín, 2006, p. 352), Member States' primary responsibility is to review the scope of the employment relationship at appropriate and regular intervals. The reviews' periodicity should be established at the national level (ILO Committee of Experts, 2020, p. 87). The Recommendation’s Preamble further argues that regular revisions are advisable because difficulties to establish an employment relationship affect not only the workers concerned, but also their communities and society at large (E. Marín, 2006, pp. 344-345). For instance, bogus self-employment can undermine the financial stability of the social security system. Moreover, since firms using bogus self-employed workers may not pay a comparable level of social security contributions, these firms can exploit an undue competitive advantage relative to competitors that do not rely on bogus self-employment. In this regard, Recommendation No. 198 affirms that “the uncertainty as to the existence of an employment relationship” must “be addressed to guarantee fair competition and effective protection of workers”; in other words, the instrument also aims at fostering a level playing field among employers. In doing so, the Recommendation acknowledges that the employment relationship is vital for the effective protection of workers, and to properly govern society more broadly.

The Recommendation consists of three main parts. A first part describes the development of a national policy on the employment relationship. A second one recommends using appropriate criteria to differentiate between employment and self-employment, and the last part stresses the need to enforce existing rules by also providing guidelines on how to do so. The national policies, very importantly, cannot be considered a one-time deal. As mentioned earlier, these policies need to be reviewed at appropriate intervals, thereby clarifying and adapting the scope of labour law, if needed. Paragraph 3 of the Recommendation also specifies that the formulation and implementation of this national policy should proceed through consultations with employers’ and workers’ most representative organisations.11 The 2020’s General Survey of the Committee of Experts on the Application of Conventions and Recommendations (CEACR), also dealing with Recommendation No. 198, has observed that social dialogue is, indeed, crucial to reach a consensus on the scope of the employment relationship at the national level (ILO Committee of Experts, 2020, p. 100).12 Paragraph 4 of the Recommendation, furthermore, stresses that the national policy should: (a) provide guidance on effectively establishing the existence of an employment relationship; (b) combat disguised employment relationships; (c) ensure that standards apply to all forms of contractual arrangements, such as those involving multiple parties; (d) ensure that it is always clear who is responsible for labour protection; (e) provide effective access to “appropriate, speedy, inexpensive, fair and efficient procedures” that can solve disputes; and (f)-(g) ensure the effective application of the laws and regulations, amongst others by providing adequate training to enforcement authorities.

The second part of the Recommendation contains more specific policy recommendations that can be used to develop employment classification mechanisms. Paragraph 8 points out that these classification mechanisms should not interfere with “true civil and commercial relationships”. This is important given that the Recommendation not only intends to tackle disguised employment,13 but arguably also provides for mechanisms and criteria to diminish the number of workers in ambiguous forms of employment.14

Paragraph 9 on the principle of primacy of facts has been considered the most critical Paragraph in the Recommendation by some commentators (e.g., R. Bignami et al., 2013, p. 11). The Paragraph clarifies that the determination of an employment relationship's existence should not be dependent on the contractual characterisation of the relationship but primarily on the facts relating to “the performance of work and the remuneration of the worker”.15 The importance of this principle should not be underestimated. Even if a report from 2013 notes that nearly all European and Latin American legal systems deem the actual implementation of the parties’ work performance to be decisive (European Labour Law Network, 2013, p. 33), this is not necessarily the case elsewhere. A report on the Republic of Korea, that was issued in 2007, details, for instance, that “the courts have been giving decisions which denied the existence of employment relationships, on the grounds of contractual arrangements or what had been agreed between the parties” (A. Yun, 2007, p. 34).16 Similarly, as Olga Chesalina argues, the Supreme Court of the Russian Federation has made an explicit reference to ILO Recommendation No. 198 in its Resolution of the Plenum No. 15 of 29 May 2018. It suggested that the whole factual situation must be taken into account when ruling about employment status. However, on many occasions, courts of lower instance have not adopted this approach; instead, they continue to base their decisions about an employment relationship on the formal contractual arrangement agreed between the parties (O. Chesalina, 2019, p. 55). In China, courts also seem to apply the principle of primacy of facts on a non-regular basis (I. Zhou, 2020, p. 34; Y. P. Chen, 2021, p. 38). Moreover, Pinto et al. (2019) also report about lobbying efforts in several States of the United States aimed at passing regulations that would undermine the primacy of facts. Therefore, even if the principle of primacy of facts seems to be broadly applied, some situations exist that warrant close attention.17

By relying on the facts, instead of the contract’s wording, as the legal basis to determine the classification of the working relationship, the classification mechanism becomes more factual in nature, rather than purely formalistic. The parties’ textual or verbal agreement on the principal’s ability to exercise control, the ability to demand the worker to work, and other contractual arrangements become less relevant under the law. It is then for the judge to decide whether the circumstances of the case at hand indicate that one of the contracting parties can indeed control, instruct, and sanction the other party, regardless of the contractual terms and conditions deployed to obtain this outcome.

Regarding the criteria or conditions that allow differentiating employment from self-employment, the International Labour Conference, in its ninety-fifth session in 2006, designated a series of elements that can be used in national classification mechanisms. Paragraph 12 of the Recommendation suggests defining clearly the conditions applied for determining the existence of an employment relationship, for example, subordination or dependence. Paragraph 13, instead, provides for a non-exhaustive list of possible indicators of the existence of such a relationship. This does not mean, however, that these indicators are mandatory, or that they are the only appropriate ones, or that they should all be present in a given work arrangement in order for it to be considered an employment relationship (N. Countouris, 2019, p. 17). Regarding the performance of work, the indicators identified in Recommendation No. 198 are: (i) whether the work is carried out according to the instructions and under the control of another party; (ii) whether the work involves the integration of the worker in the organisation of the enterprise; (iii) whether the work is performed solely or mainly for the benefit of another person; (iv) whether the work must be carried out personally by the worker; (v) whether the work is carried out within specific working hours or at a workplace specified or agreed by the party requesting the work; (vi) whether the work is of a particular duration and has a certain continuity; (vii) whether the work requires the worker's availability; or (viii) whether it involves the provision of tools, materials and machinery by the party requesting the work. Concerning the remuneration of the worker, the Recommendation indicates: (i) the periodic payment of remuneration to the worker might matter; (ii) the fact that such remuneration constitutes the worker's sole or principal source of income could be deemed important; (iii) the provision of payment in kind, such as food, lodging or transport can be taken into account. Other relevant indicators are the entitlement to weekly rest and annual holidays, the payment of travel expenses to carry out the work, and the absence of financial risk for the worker. The determination of what elements are important or even determinative is a matter that should be settled through a continually evolving national policy. In this respect, the CEACR has suggested that the fact that, for instance, tools and materials are provided for by the worker should not be a determinative indicator. The extent to which the worker is integrated into the business organisation, on the other hand, would require “careful consideration” (ILO Committee of Experts, 2020, p. 111).18 It is also worth noting that, according to the Committee, “[c]urrent indicators may no longer be useful in determining the existence of future employment relationships. Member States should therefore consider the need to establish new criteria and disregard existing criteria when they are no longer useful. This would be in accordance with Paragraph 1 and the need to periodically review, clarify and adapt the scope of relevant laws to guarantee the effective protection of workers in an employment relationship” (ILO Committee of Experts, 2020, p. 113).

Paragraph 11, furthermore, points out that a broad range of means should be considered in the determination of the existence of an employment relationship. The same Paragraph also stresses that to facilitate the determination of an employment relationship, States can devise legal presumptions where one or more relevant indicators are present. This would shift the burden of proof and might help workers to more easily claim their employment rights. Several examples of legal presumptions exist in national practices (ILO, 2016, p. 264-265). Paragraph 11 also mentions that States should determine that workers with specific characteristics, in general, or in a sector, must be deemed to be either employed or self-employed.19 The CEACR has recently stated that these kinds of presumptions are “crucial to counterbalance the unequal bargaining power of the parties and as a consequence of the principle in dubio pro operario which is fundamental in labour law” (ILO Committee of Experts, 2020, p. 104).

Recommendation No. 198 also addresses the problematic enforcement of laws and regulations. This is a crucial concern. Research in the European Union has, for instance, observed that EU Member States seem to deal with undeclared work actively, but “that actions to tackle [bogus self-employment] in the eight case study countries [analysed] are, by comparison, less developed and that public authorities (including enforcement bodies) in some countries regard [bogus self-employment] as a relatively low priority” (J. Heyes and T. Hastings, 2017, p. 52). According to some scholars, the issue of a lack of enforcement is all the more important with regard to the platform economy, where it has been challenging to distinguish employed from self-employed workers (D. Minerva and R. Stefanov, 2018, p. 13).

Recommendation No. 198 approaches the issue of enforcement from a broad perspective. Paragraph 17 suggests that Members should remove private individuals’ incentives to disguise an employment relationship. Most notably, tax and social law can provide the parties with incentives to use self-employment.20 Therefore, labour inspectorates should collaborate with the social security administration and tax authorities to ensure respect for laws and regulations, as suggested by Paragraph 16.21 In addition to these preventive steps, enforcement programmes and processes should be regularly monitored according to Paragraph 15. Lastly, Paragraph 22 stresses that “Members should establish specific national mechanisms in order to ensure that employment relationships can be effectively identified within the framework of the transnational provision of services. Consideration should be given to developing systematic contact and exchange of information on the subject with other States.” Arguably, this Paragraph of the Recommendation could provide a reference to pursue the ILO Global Commission on the Future of Work’s proposal to establish an international governance system for digital labour platforms (Global Commission, 2019, p. 44). Paragraph 22 was inserted because the EU Member States, in particular, recognised the difficulties “in establishing who is considered a worker, what rights the worker has and who is accountable for those rights” in the framework of transnational services.22 However, the Employers’ group insisted on limiting the Paragraph’s recommendation to “national mechanisms”. This amendment was eventually accepted, even if it “unnecessarily limited flexibility” according to the Worker Vice-Chairperson.23 Despite the reference to “national mechanisms”, the CEACR still considers Paragraphs 7 and 22 “of great relevance” for “new and emerging forms of the organisation of work that involve multiple processes with different actors across borders, including digital platform workers” (ILO Committee of Experts, 2020, p. 99). Paragraph 7 calls upon Members to, amongst others, consider concluding bilateral agreements to prevent abuse and fraudulent practices when workers are recruited in one country to work in another.

Even if this summary of the Recommendation has illustrated that the instrument was in many ways forward-thinking, some very important aspects related to the protection of certain categories of vulnerable workers are not included in the Recommendation; or if they are, they are only partially covered. As discussed, the possibility to identify a broader scope of application for at least some labour protections was abandoned together with the draft instrument on Contract Labour. Furthermore, the Recommendation mainly focuses on the scope of ‘employment law’, which is principally considered to deal with the rights and obligations of employers and workers. Aside from a general reference to Fundamental Principles and Rights at Work and Decent Work, and thus also to freedom of association and the right to collective bargaining, in the Preamble, the collective labour rights are scarcely mentioned in the text of the Recommendation. This, however, can also depend on the fact that the ILO Supervisory Bodies consistently consider the rights and principles relevant to freedom of association and the right to collective bargaining, applicable to all workers, regardless of their employment status.24 Additionally, the Recommendation also demands to pay particular attention to workers who are more likely to be affected by uncertainty as to the existence of an employment relationship, pointing amongst others to migrant workers and workers in the informal economy (Paragraph 5). It does not, however, specifically mention workers who operate in industries with high levels of employment and social security fraud, or workers who are engaged in occupations that are at a higher risk of being misclassified as self-employed workers, and thus deprived of a wide range of labour protection (e.g. journalists, sales representatives).25

Another topic that remains unaddressed are the so-called ‘intermediate or hybrid categories’ of quasi-subordinate or semi-dependent workers, which are closely related to the proper application of employment protection. The creation of a hybrid category may, indeed, affect the personal scope of the traditional employment relationship, as some workers that, absent a hybrid category, would have been included in the notion of ‘employee’ may find themselves excluded by this notion when such a category exists at the national level. Moreover, hybrid categories have an unquestionable impact on the effective enforcement of employment, labour and social security regulation (M. Cherry and A. Aloisi, 2017; OECD, 2019, pp. 143-145). Intermediate categories are, in other words, inextricably linked to the subject matter of the Recommendation. The CEACR seems to have acknowledged this recently. The Committee mentioned that a “grey zone” between employment and self-employment has always been present, but has become more prominent due to “recent changes in the organisation of work, together with technological developments”. In response, the Experts point out, certain countries have turned the Recommendation’s “binary approach” into a ‘“modified binary divide” model’ with intermediary categories (ILO Committee of Experts, 2020, pp. 117-118).

Finally, the question of “who is the employer?” also remains largely unaddressed. The Recommendation does stipulate in Paragraph 4 (c) that national policy should at least “ensure that standards applicable to all forms of contractual arrangements establish who is responsible for the protection contained therein” and “ensure standards applicable to all forms of contractual arrangements, including those involving multiple parties, so that employed workers have the protection they are due”. Nevertheless, unlike the draft Recommendation on contract labour, which proposed making the user enterprises responsible if the subcontractors fail to perform their obligations, and vice-versa (ILO, 1998a, pp. 13-15), Recommendation No. 198 does not provide specific instructions on these matters. This is arguably a loophole since the identification of the employer in a work arrangement is one of the essential issues in determining the existence of an employment relationship (J. Prassl, 2015).26

In concluding this section, it is also essential to notice how the employment relationship's continuing relevance as a social and legal institution has been repeatedly affirmed at the ILO level. For instance, the Preamble of the 2008 Social Justice Declaration recognises the importance of the employment relationship as a means to provide legal protection. The 2019 Centenary Declaration similarly reaffirmed “the continued relevance of the employment relationship as a means of providing certainty and legal protection to workers”. The employment relationship will, therefore, most likely remain a cornerstone in the International Labour Organisation’s agenda for the foreseeable future. Concomitantly, the Employment Relationship Recommendation, 2006 (No. 198), will probably remain at the centre of the Organisation’s normative activity as well.

2.1. Self-employed workers’ provisions

Platform operators unilaterally determine the terms and conditions of their engagement with clients and workers. These terms and conditions most commonly classify platform workers as “self-employed”, or “independent contractors”. Platforms also regularly suggest that the relationship of the worker with the client remains one of self-employment. Some examples of such terms and conditions are provided below. We have emphasized certain aspects in bold.

• Amazon Mechanical Turk: “Workers perform Tasks for Requesters in their personal capacity as an independent contractor and not as an employee of a Requester or Amazon Mechanical Turk or our affiliates. As a Worker, you agree that: (i) you are responsible for and will comply with all applicable laws and registration requirements, including those applicable to independent contractors and maximum working hours regulations; (ii) this Agreement does not create an association, joint venture, partnership, franchise, or employer/employee relationship between you and Requesters, or you and Amazon Mechanical Turk or our affiliates; (iii) you will not represent yourself as an employee or agent of a Requester or Amazon Mechanical Turk or our affiliates; [...] As a Requester, you will not engage a Worker in any way that may jeopardise that Worker's status as an independent contractor performing Tasks for you.”27

• Doordash: “CONTRACTOR represents that he/she operates an independently established enterprise that provides delivery services, and that he/she satisfies all legal requirements necessary to perform the services contemplated by this Agreement. As an independent contractor/enterprise, CONTRACTOR shall be solely responsible for determining how to operate his/her business and how to perform the Contracted Services. [...] The parties acknowledge and agree that this Agreement is between two co-equal, independent business enterprises that are separately owned and operated. The parties intend this Agreement to create the relationship of principal and independent contractor and not that of employer and employee. The parties are not employees, agents, joint venturers, or partners of each other for any purpose. Neither party shall have the right to bind the other by contract or otherwise except as specifically provided in this Agreement.”28

• Ola: “During the Term of this Agreement, the Transport Service Provider shall operate as and have the status of an independent contractor and shall not act as, be or construed to be an agent or employee of OLA. The relationship between the Parties is on a principal-to-principal basis, and none of the provisions of this Agreement shall be interpreted as creating the relationship of employer and employee between the Transport Service Provider and OLA at any time, under any circumstances or for any purpose. [...].”29

• Taskrabbit: “Taskers are independent contractors of clients and not employees, partners, representatives, agents, joint venturers, independent contractors or franchisees of Taskrabbit. Taskrabbit does not perform tasks and does not employ individuals to perform tasks. users hereby acknowledge that company does not supervise, direct, control or monitor a tasker’s work and expressly disclaims (to the extent permitted by law) any responsibility and liability for the work performed and the tasks in any manner, including but not limited to a warranty or condition of good and workmanlike services, warranty or condition of quality or fitness for a particular purpose, or compliance with any law, regulation, or code.”30

• Handy: “Service Professional is an independent contractor and has not been engaged by Handy to perform services on Handy's behalf. Rather, Service Professional has entered into this Agreement for the purpose of having access to the Handy Platform, in exchange for which it pays Handy a fee, as described herein. Service Professional represents that he or she is customarily engaged in an independently established trade, occupation, profession and/or business offering the Services to the general public and/or Service Professional represents that he or she maintains a principal place of business in connection with Service Professional’s trade, occupation, profession and/or business that is eligible for a business deduction for federal income tax purposes. [...].”31

• Instacart: “You enter into this Agreement as an independent contractor with a business relationship between you and Instacart. It is understood that in agreeing to provide Services under this Agreement, the Contractor shall be acting and shall act at all times as an independent contractor, and not as an employee of Instacart for any purpose whatsoever, including without limitation, for purposes relating to taxes, payments required by statute, or any other withholdings or remittances to any governmental agency or authority. [...] You further acknowledge that this Agreement does not create any employer-employee relationship between a third party retailer and yourself, and that you are not entitled to any benefits, including but not limited to Workers’ Compensation coverage, afforded to any employees of a third party retailer.”32

Other platform operators try to minimize the importance of their contribution to the overall transaction:

• Care.com: “Care.com does not introduce or supply Carers to Care Seekers, nor do we select or propose specific Carers to Care Seekers or Care Seekers to Carers. Instead, we offer an online forum that among other things enables Care Seekers and Carers to interact regarding potential work opportunities.”33

• Urbansitter: “No Joint Venture or Partnership. Nothing in these Terms of Service may be construed as making either party the partner, joint venturer, agent, legal representative, employer, contractor or employee of the other. UrbanSitter is not an employment service or agency, and does not serve as an employer or referral source for any User.”34

• Sittercity: “No Joint Venture. You acknowledge that you are not legally affiliated with Sittercity in any way, and no independent contractor, partnership, joint venture, employer-employee or franchiser-franchisee relationship is intended or created by your use of the Service or these Terms of Use. As such, you shall not have, or hold out to any third party as having, any authority to make any statements, representations or commitments of any kind, or to take any action, that shall be binding on Sittercity, except as provided herein or authorised in writing by Sittercity. Sittercity is not an employment service or agency, does not serve as an employer of Users and does not recruit Users for employment, secure employment for Users or evaluate or test Users for employment purposes.”35

Certain platforms have also included clauses that seem to shed potential responsibilities arising from the transaction to the other parties (i.e. final clients and workers). These responsibilities may be linked to the possible reclassification of workers into employees or to compliance with existing employment laws. Some examples are:

• Bizzby: “14 STATUS AND TAX

• 14.1. As a Service Provider, you acknowledge and agree that you are an independent sole-trader or corporate Service Provider and are not engaged as an employee, contractor or agent of BIZZBY and you remain solely responsible for:

• 14.1.1 your own Tax, VAT and employment affairs (and that required to be paid for your employees and subcontractors (where applicable); and

• 14.1.2 the quality and outcome of the work and performance of the Jobs, and you will indemnify and hold BIZZBY entirely harmless from any claims from (i) [HM Revenue and Customs] or any other statutory body, or (ii) your employees or subcontractors, arising from a breach of the above.”36

• Care.com: “Each Care Seeker is also responsible for complying with all applicable employment, anti-discrimination and other laws in connection with any relationship they establish with a Carer. Among other things, Care Seekers must verify that the Carer they select is legally entitled to live and work in the UK, and perform any other verification or other checks for the position they are hiring, such as identity, qualification, background and reference checks. In addition, Care Seekers must comply with all applicable anti-discrimination legislation including but not limited to the Equality Act 2010”.37

• Taskrabbit: “As set forth in section 1, taskrabbit does not perform tasks and does not employ individuals to perform tasks. Each user assumes all liability for proper classification of such user’s workers based on applicable legal guidelines. If a client, you indemnify and hold taskrabbit and affiliates harmless, and if a tasker, you fully and finally release taskrabbit and affiliates, from all liabilities, claims, causes of action, demands, damages, losses, fines, penalties or other costs or expenses that taskers or assistants may incur or become entitled to, whether under contract, common law, civil law, statute or otherwise, in respect of tasks or service agreements or the use of the taskrabbit platform, including with respect to misclassification of taskers and assistants and the termination or cessation of any tasks, service agreements, this agreement or the use of the taskrabbit platform.”38

• Handy: “5. RELATIONSHIP OF THE PARTIES Service Professional is an independent contractor and has not been engaged by Handy to perform services on Handy's behalf. [...] Handy will not be responsible for withholding or paying any income, payroll, Social Security, or other federal, state, or local taxes, making any insurance contributions, including unemployment or disability, or obtaining workers' compensation insurance on Service Professional's behalf. Service Professional shall be responsible for, and shall indemnify and hold Handy harmless for any claims, suits, or actions related to this provision, including any such claims brought by Service Professional or by any third party with respect to any claims for taxes or contributions, including penalties and interest.”39

Some platforms have also altered the provisions concerning their workers’ employment status over the years. For instance, while some years ago, the terms and conditions of the Amazon Mechanical Turk platform (AMT) had a provision that seemed to warn clients for potential reclassification of workers as employees, this no longer seems to be the case.40 Nowadays, the AMT asks users to not engage workers in a way that might jeopardise their classification as independent contractor (see supra). Other platforms, such as Deliveroo, have decided to no longer use fixed-term employees and now engage independent contractors.41

2.2. “Employees’ provisions”

Some platforms seem to allow users to engage platform workers as employees of a payroll service company or a third-party staffing vendor. 

• Wonolo: “Upon accepting a Wonolo Request or an Open Request that is classified as Payroll with Customer through the Application (each a “Payrolled Wonolo Engagement”), Wonoloer agrees and understands that Frontline will choose the payroll service company (the “Payroll Company”) to be the employer of record, meaning the Wonoloer will be an employee of Payroll Company (a “Payroll Employee”) solely for Payrolled Wonolo Engagements. All other Wonolo Engagements will be on an independent contractor basis, as outlined in the Terms of Service. Frontline will instruct Payroll Company to assign Wonoloer to work for Customer, and Customer is responsible for supervising Wonoloer.”42

• Upwork: “When a Client uses Upwork Payroll, which is described on the Site here (“Upwork Payroll”), a third-party staffing vendor will employ the Freelancer (the “Staffing Provider”). Freelancer (if accepted for employment as described below) will become an employee of the Staffing Provider. The Staffing Provider will assign Freelancer to work for Client, and Client will be responsible for supervising Freelancer. When, and only if, a Freelancer has been accepted for employment by the Staffing Provider and assigned to Client, Freelancer becomes a “Payroll Employee” for purposes of this Agreement, but also remains a Freelancer under the Terms of Service. Your ability to use Upwork Payroll may depend on certain factors, including, without limitation, the location of the Freelancer, the estimated length of the engagement, the wage to be paid, and the nature of the work to be performed. A request to use Upwork Payroll may be rejected for any lawful reason.”43

Deliveroo used to make recourse to workers hired as employees by a cooperative that staffed the riders to the platform in Belgium.44 The platform, however, unilaterally decided to discontinue this scheme as of February 2018.

Hilfr, a Danish-owned platform for cleaning in households, most notably allows the workers to work either as freelancers or as employees (A. Ilsøe and K. Jesnes, 2020 pp. 53-59). A collective agreement concluded between this platform and the union 3F stipulates the following.

• Hilfr: “The collective agreement covers cleaning assistants who perform cleaning work in private households facilitated by the digital platform Hilfr ApS. (hereinafter Hilfr). The collective agreement covers cleaning assistants who are employees, not freelancers. [...] Freelancers automatically obtain employee status after 100 hours’ work via the platform and are subsequently covered by this collective agreement. [...] Freelancers who wish to transfer their status from freelancer to employee before having worked 100 hours must notify Hilfr of this. [...] Freelancers who wish to remain freelancers after 100 hours’ work facilitated by the platform must inform Hilfr of this decision well in advance of the expiry of the 100 hours.”45

Another example of a platform that allows its workers some choice can be found in the UK. DPD, which is a delivery business for parcel-delivery, offers workers three choices. According to DPD’s brochure, drivers can be employed drivers, self-employed owner-driver workers (ODWs) or self-employed owner-driver franchisees (ODFs). The company offers sick and holiday pay to all its drivers.46 Moreover, the enterprise has abolished its £150 daily fines for missing work, as a response to the death of a driver with diabetes who arguably collapsed because the company charged for attending a medical appointment (R. Booth, 2018). Hermes, another parcel delivery company that also operates in the UK, reached an agreement with GMB, one of the UK’s largest trade unions. The agreement gives couriers the option to become ‘self-employed plus’, which provides several benefits, such as holiday pay and guaranteed earnings (GMB Union, 2019). Furthermore, Voocali, which provides translation services in Denmark, also concluded a collective agreement with the Union of Commercial and Clerical Employees, which falls apart into, on the one hand, “an accession agreement so that future employees of Voocali get decent pay- and working conditions” and on the other, “an agreement that covers all those who are not covered by the collective agreement – i.e. the freelancers” (Io, 2018).47 The latter apparently covers the pay and employment terms for freelancers (A. Ilsøe et al., 2020, p. 74).

Several other platforms have at least engaged some platform workers as employees at one stage or another. A collective agreement in Sweden has, for instance, been signed between Bzzt, a company that uses environmentally-friendly electric vehicles, and the Swedish Transport Workers’ Union. As a result, Bzzt drivers are covered by an industry-wide collective bargaining agreement that gives them access to the same standards as traditional taxi drivers. They are offered marginal part-time employment contracts (H. Johnston and C. Land-Kazlauskas, 2018, p. 30). The employees are entitled to monthly or weekly wages, receive a minimum wage, pension, holiday pay and sick pay. Arbitration is provided to solve disagreements (A. Ilsøe et al., 2020, p. 74).

YouGenio, an Italian platform providing domestic work, such as cleaning, babysitting, hydraulic and electrical services, hires its workers under an employment contract in its terms of service; where it seems to indicate that the workers can be either employees of the platform, employees of a third party, or independent contractors.48

Takeaway promises “a proper employment contract” on its recruitment page.49 The terms and conditions of service, however, are not openly consultable online. In Norway, Foodora has agreed to classify its workers as employees since its establishment in the country (B. Lindahl, 2019). The company employs its workers mostly on a part-time basis. Foodora employees in Norway have also, after a five-week strike, reached a collective agreement to improve their working conditions, in particular, about payment for actual working time and reimbursement for equipment (Svensson-stiftelsen, 2019). In this respect, Ilsøe and Jesnes note: “Unlike Deliveroo and Wolt couriers (and Foodora couriers in other countries), who are paid per delivery and engaged as self-employed, Foodora couriers in Norway are hired on marginal part-time contracts (10 hours per week)” (2020, p. 60). Interviews carried out in Germany have, furthermore, shown that “only temporary contracts (mainly with a one-year duration) are used at Foodora, whereas Deliveroo makes use of both temporary employment contracts (with a duration of about six months) and self-employment” (T. Haipeter et al., 2020, p. 12).

Helpling explicitly engages employees in Switzerland. The respective terms and conditions state that “Helpling offers cleaning activities performed by Helpling employees. [...] Customers agree that they do not independently contact, hire or otherwise maintain Helpling's employees beyond the scope of Helpling's services. In particular, the customer may not hire Helpling employees. This is prohibited even after termination of the contract between Helpling and the customer for one year.”50 It is worth mentioning in this respect that Helping defines its workers as independent contractors in most of the other European legal systems in which it operates. In the Netherlands, however, it does allow its workers to be employed as well. Differently from the Swiss context, where the employment relationship is established between the worker and platform operator, its Dutch cleaners are considered employees of the household, according to a court ruling. The judgment also considers Helpling a private employment agency under national law.51 This is not a unique situation. The user clients of Batmaid, a cleaning platform operating in Switzerland, also conclude employment contracts with the workers. The platform’s help page clarifies that “[o]ur clients are the only employers of the trusted cleaning agents. The Batmaid platform just facilitates the selection of a relevant trusted cleaner available to your doorstep and takes care of all the paperwork, legal declaration to the AHV (AVS), social security insurance payments and issues monthly pay slips and annual wage statements.”52

Lastly, some platforms perform their services by voluntarily adopting the legal status of a temporary work agency. This has been the case, for instance, for Ploy in Belgium and the Netherlands, which is owned by Randstad, a private employment agency,53 and Zenjob in Germany.54

3.1. Introduction to legislative measures on platform work

A joint statement by the 2000’s Meeting of Experts on Workers in Situations Needing Protection noted that “the global phenomenon of transformation in the nature of work has resulted in situations in which the legal scope of the employment relationship (which determines whether or not workers are entitled to be protected by labour legislation) does not accord with the realities of working relationships. This has resulted in a tendency whereby workers who should be protected by labour and employment law are not receiving that protection in fact or in law.”55 This observation is valid to this day. In a recent General Survey of the CEACR, amongst others, Cambodia, Germany, Latvia, Mauritius and Peru noted that they found it challenging to ensure labour protection for platform workers (ILO Committee of Experts, 2020, p. 90). Also, Norway expressed this sentiment, even though most experts in the Norwegian Committee on the Sharing Economy found it unnecessary to develop a new category of workers with limited protections, nor to establish a legislative committee to revise the conceptual apparatus of the Working Environment Act. A new, intermediate category may, for instance, enable the circumvention of ordinary employment protection (Departementenes sikkerhets- og ser viceorganisasjon, 2017, pp. 11 and 75).

Therefore, it is interesting to note how multiple other legal systems did, in fact, make changes to their laws and did so with radically different aims. This Section will discuss these legislative developments. It differentiates between the various legislative or regulatory efforts by discussing in Section 3.2., a first category of legislation that regulates platform work, as such, and in Section 3.3., a range of legislation that alters the criteria or mechanisms meant to differentiate between employees and self-employed workers on the labour market as a whole.

3.2. Platform-specific regulatory measures

“Platform-specific” measures may serve very diverse regulatory aims. In the United States, for instance, a policy brief of the National Employment Law Project noted that “[o]ver the 2018 legislative session, nearly identical bills have been introduced in states including Alabama, California, Colorado, Florida, Georgia, Indiana, Iowa, Kentucky, Tennessee and Utah that deem all workers on so-called “marketplace platforms,” (such as Uber and Handy) independent businesses, and not the employees of the companies” (R. Smith, 2018, p. 1). Some of these bills have by now made it into law. For example, Iowa has a law on the books that defines ‘marketplace contractors’ and classifies them as independent contractors for all purposes under state or local law.56 Utah similarly has a “Service Marketplace Platforms Act” with a presumption that a ‘building service contractor’ is an independent contractor.57 These acts have a different scope but share a similar aim, namely an attempt to exclude the existence of an employment relationship. Moreover, administrative authorities have taken similar measures. For instance, the Texas Workforce Commission, which is the agency responsible for administering unemployment benefits and assessing unemployment taxes, has adopted a rule that stipulates that certain workers who provide services via app-based businesses and websites cannot be considered “employees” for unemployment insurance purposes (E. Douglas, 2019).58 Even if some States have resisted this type of bills, legislation to “carve-out” platform workers from employment and labour regulation passed in many other States, including Arizona, Florida, and Indiana (M. Pinto et al., 2019).

Platform workers have also been statutorily classified as self-employed in Belgium, at least to a certain extent. La loi relative à la relance économique et au renforcement de la cohésion sociale of 2018 has framed platform work as one of the three forms of work of which the generated income is considered “auxiliary” and is tax-exempt.59 As long as the worker does not earn more than 6.340 euro a year – in addition to a monthly threshold – through recognised platforms, this income is exempted from regular taxes and social security contributions. The law also specifies the conditions to fall under this beneficial tax status.60 None of these conditions provides the worker with any social protections. Nor does the certification of the recognised platforms by the government require the platform to make any social commitments.61 Instead, according to some scholars, legislation has embedded this form of platform work through recognised platforms, in a sort of ‘social law free zone’ (Y. Stevens and J. Put, 2018, p. 147).

Because of this lack of social protection, and the low fiscal and social security charges involved, both trade unions and employers’ associations lodged a case at the Constitutional Court against all three the schemes of the 2018 Act. The Court declared the Act unconstitutional. All three forms of auxiliary income, including the one for platform work, were deemed to violate the Constitution. Regarding platform work, according to the Court, the legislature had made an “unsubstantiated assumption” by branding, amongst others, platform workers’ income as “auxiliary”. Even if this assumption were well-founded, the Court noted, it would still not justify a different treatment compared to the one applying to the income necessary to deal with living expenses.

Furthermore, while uncertainty about the classification of these platform work relationships may perhaps warrant a separate employment category, in the view of the Court, it does not justify a complete exemption from employment legislation, from the social security system and tax obligations. The Court also ruled that the effects of the annulled 2018 Act are maintained until 31 December 2020, not to harm the workers affected.62 Accordingly, from 2021 onwards, platform workers will again provide their work under the previous Act, 2016’s loi De Croo. This Act already contained beneficial tax provisions for platform workers, but did not explicitly exclude them from employment legislation.63 As a result, unless the legislature advances an amendment, Belgian employment legislation will no longer contain any sections that prevent courts from designating a platform worker regulated under the loi De Croo as an employee.

Apart from these examples, several countries and institutions have recently passed various measures to protect certain platform workers. Regulation 2019/1150 of the European Parliament and of the Council of 20 June 2019 on promoting fairness and transparency for business users of online intermediation services will, for instance, provide certain self-employed platform workers with useful protections. It intends to provide “business users” that use “online intermediation services” to offer goods or services to consumers with transparent conditions. It tries to make sure that they are treated fairly and aims to provide effective redress vis. the “intermediation service provider”.64

An important limitation of this instrument’s scope, however, is that online intermediation services are defined as information society services that “allow business users to offer goods or services to consumers, with a view to facilitating the initiating of direct transactions between those business users and consumers”, and which “are provided to business users on the basis of contractual relationships between the provider of those services and business users which offer goods or services to consumers”. This is important because, according to the Court of Justice of the European Union, for instance, Uber is not an information society service provider,65 while a platform like Airbnb does qualify as such.66 The question of whether a platform operator provides services other than digital intermediation is, therefore, a crucial one. According to the Court of Justice of the EU, Uber provides services in the field of transport, rather than only acting as an intermediary. Hence, the Member States can regulate this platform’s activities at the national level. Uber is not just an information society service provider. Therefore, it arguably does not necessarily have to ensure that its terms and conditions comply with the standards promoting fairness and transparency mandated in the EU Regulation.

In the EU, France has also granted certain labour protections to some platform workers. Act No. 2016-1088 of 8 August 2016 introduced a protective scheme for those self-employed platform workers who provide their services on a platform that determines the service’s characteristics and sets each task’s price (e.g. a ride, a delivery, or cleaning task).67 The Act requests platforms to assume their social responsibility and to, for instance, provide pay for insurances against workers’ occupational accidents, if the worker’s pay exceeds a certain threshold set by Decree.68 The same condition applies to the worker’s right to vocational training. Sections L7342-5 and L7342-6 of the Labour Code also mention a right to form or join a trade union and engage in concerted refusals to provide their services. Platforms are prohibited from retaliating when workers defend their professional claims.

While this regulatory scheme came into effect for self-employed platform workers, certain courts have ruled on platform workers’ employment status. As we discuss below, French higher courts have classified some platform workers as employees. In 2018, however, a bill concerning the loi d’orientation des mobilités aimed to regulate platform work in the transport industry. This Mobility Act was adopted, and it enshrines the platforms’ social responsibility towards workers who provide transportation of passengers and delivery of goods using a 2- or 3-wheel vehicle. This is done through the voluntary and unilateral adoption of charters, which define both parties’ rights and obligations.69 According to the original bill, this would have made it more difficult for the platform workers concerned to claim employment status. The bill stipulated that once a charter was approved between the platform and administrative authorities, the rights and obligations included in the charter could not have been considered indicators of legal subordination between the platform and the relevant workers.70 The Conseil constitutionnel deemed this mechanism unconstitutional on the grounds that the legislature would have allowed platform operators to unilaterally state the features of their relationship with their workers through a charter, something that would have prevented the courts from carrying out their own assessment when ruling on a reclassification lawsuit. Platforms would thus have been able to set rules which are, according to the Conseil constitutionnel, to be set by law.71 Section 34 of the French Constitution clarifies that statutes shall lay down the basic principles of, amongst others, employment law and trade union law.72 One of the fundamental principles of labour law, according to the Conseil, is the determination of its scope. Therefore, the private parties’ charter cannot condition what the labour courts can and cannot consider as indicators of an employment relationship. Apart from this mechanism, deemed unconstitutional, the Mobility Act inserted new regulations in the code du travail, mandating specific provisions for the social charter of certain platforms that engage in transportation and delivery of goods by 2- or 3-wheeled vehicles.73

In Colombia, a bill was recently introduced to regulate economically dependent platform work.74 This bill considers the economically dependent platform worker providing services to a client through a mobile app or platform difficult to fit into the existing dichotomy of labour relations in Colombia.75 Therefore, the draft law suggests creating an intermediate category for economically dependent platform workers. The relationship between these ‘platform workers’ and ‘the digital intermediation company’ is labelled as a “substantive” one, instead of the employment and civil law relationships. Furthermore, notwithstanding the explicit denial of an employment relationship, the explanatory memorandum highlights the importance of the principle of the primacy of facts.76 The creation of the new legal category for economically dependent platform workers is meant to establish certain protections, some of which consider the peculiar nature of services provided by platform workers. Some examples are the opportunity for portable ratings, or the prohibition to allocate customers to the worker on a compulsory basis. The bill also provides for the enjoyment of collective labour rights, such as freedom of association and the right to bargain collectively. The digital intermediation companies must facilitate the exercise of these rights, e.g. to provide the contacts of other workers, and not sanction workers with deactivation due to their labour claims or disagreements resulting from their relationship with the company. Furthermore, the bill also stipulates some social security protections. It proposes to make the contribution to the health and pension system compulsory and in equivalent parts by both the platform and the worker. Insurance against accidents is also provided not only for platform workers, but also for service users. In the same timeframe, another bill on virtual labour was proposed.77 It is worth noting that this bill explicitly sets out not to apply to platform work,78 as it aims to create a new type of contract and employment relationship, concretely ‘virtual labour’, where all labour relations are carried out online using technology.

Also, in Argentina, a bill has been advanced to protect workers on digital on-demand platforms. The bill provides several protections, such as a maximum 48 hours workweek, a daily rest time of at least 12 hours, minimum guaranteed remuneration and a day of holiday for every 120 hours worked. It also includes mechanisms to calculate the compensation in case of unjustified dismissal.79

In Italy, new provisions relevant to platform work were recently introduced. In 2019, the lawmakers amended80 a provision introduced in 2015 and aimed to expand the scope of employment protection beyond the employment relationship to all workers providing work organised by another party, including via a platform.81 As a result, employment and labour protection would apply to these platform workers, unless a collective agreement provides otherwise (A. Aloisi and V. De Stefano, 2020).82 The 2019 amendment meant to clarify the scope and operation of the 2015 provision, whose applicability to platform work had given rise to controversies among courts.83 However, the Italian Supreme Court's judgment from 2020 confirmed that the 2015 reform already applied to workers whose work is organised by a platform.84 On the basis of this legislation, in February 2021 the Italian labour inspectorate ordered the major food-delivery platforms to apply employment and labour protection to up to 60,000 riders in Italy and mentioned the possibility of applying hefty sanctions for alleged irregularities (E. Parodi, 2021). Furthermore, a residual regulation was inserted in the same legislative decree,85 which was dedicated exclusively to self-employed delivery couriers, in the unlikely event that they would not fall within the personal scope of application of the 2015 provision, as interpreted by the Supreme Court and as amended in 2019.86 This residual regulation concerns minimum pay and collective bargaining, the need for a written contract and mandatory insurance against accidents at work and occupational diseases.87 Importantly, at the end of 2020, the Tribunal of Palermo88 reclassified a rider of the Glovo platform as a subordinate employee. In this judgment, a court found a standard employment relationship between a worker and a platform for the first time in Italy. According to the court, the rider in question was entitled to a full-time and permanent contract.

Some regional initiatives are also relevant. In Italy, Lazio became the first region in the country to approve a law on the labour protection of ‘digital workers’ in 2019.89 The personal scope of application of this law extends to all workers whose work is organised via an app. In addition to regulation regarding health and safety issues, the law calls for collective bargaining to establish the levels of basic pay and indemnities. Furthermore, the law provides for the creation of a Digital Work Portal for the registration of both platforms and workers, and the establishment of a Regional Digital Labour Council, in charge, among other things, to monitor the conditions of digital work in the Region and to facilitate social dialogue between platforms, workers and the social partners.

Some countries have taken steps to include some platform workers within the scope of the employment protection. In Portugal, for instance, a so-called “Uber Law”90 came into force to set out the rules for the individual and remunerated transport of passengers. It establishes a legal regime for electronic platforms providing this mode of transport.91 A presumption of employment applies to the contract between the platform and the driver (see also ILO Committee of Experts, 2020, p. 141). Section 13 of the law also stipulates that, regardless of the number of platforms a driver works for, they may not operate vehicles for more than ten hours within a 24-hour period. Platform operators must implement mechanisms designed to achieve this end. Platforms must also provide certain information to drivers, and devise a mechanism for complaints (section 19). The law also requires the computer system to track the driver’s working time, including rest time (section 20).

Furthermore, social dialogue in Spain led to an agreement about a forthcoming “riders law” (A. F. Varela, 2021). Similarly to Portugal’s Uber law, at the heart of this law there lies a rebuttable legal presumption of an employment relationship for food delivery riders in Spain. At the same time, according to some commentators, this specific legal presumption might be more difficult to prove than the general legal presumption of an employment relationship (A. Todolí, 2021). Four conditions need to be fulfilled for activation of this specific presumption: (i) the provision of services by one person; (ii) the delivery of goods to final consumers; (iii) the direct or implicit exercise of the employer’s management powers through a digital platform; (iv) and the use of an algorithm to manage the service or to determine the working conditions. Therefore, the court will determine the applicability of this legal presumption, ruling on a case-by-case basis.

Other legal systems might have laws, regulations, or collective agreements that specifically apply to (certain) platform workers, but only if they are already considered as employees. In Austria, for instance, a collective agreement has been concluded between the Chamber of Commerce (employers’ association) and the Austrian Union of Service workers, supposedly covering all bicycle delivery services (H. Johnston, 2020, p. 32). As such, this agreement has a scope of application similar to a statutory one, with an erga omnes effect. The agreement “does not touch on the issue of whether bike couriers should be qualified as workers or contractors” (J. Widner, 2019), but it does regulate the employment relationship of bike couriers from 1 January 2020 onwards. Although, in other words, this agreement does not directly influence the classification of the relationship between a worker and a platform, it does provide for a minimum wage, holiday allowances, expense allowance for the use of the bicycle and so forth, for couriers that classify as employees under the current criteria used to determine the existence of employment status (Ibid.).

3.3. Legislation on the employment relationship

Many countries have legislation in force that tries to enhance the clarity of the framework and conditions that differentiate between employment and self-employment, or have taken measures to counter misclassification, as suggested by the Employment Relationship Recommendation, 2006 (No. 198). These changes are not less important than the “platform-specific” measures, because as the CEACR also considers, “the use of technological means to distribute tasks to an indeterminate workforce cannot justify these activities being considered forms of work separate from the rest of the labour market” (ILO Committee of Experts, 2020, p. 142). Platform work is not an idiosyncratic phenomenon (V. De Stefano, 2016a). Moreover, many Member States realise the importance of revising existing provisions and procedures to deal with bogus self-employment, including in the context of platform work.

In relation to the framework and conditions that settle employment status, it is clear that the courts play a paramount role in distinguishing between employees and self-employed workers. They are called upon to apply abstract legal criteria to the cases at hand. This is, for instance, the case in Switzerland. A Chambre administrative de la Cour de justice in Geneva has acknowledged that Uber Eats’ delivery couriers work on the basis of an employment relationship. Uber has appealed the decision, taking it to the Federal Court. Therefore, the outcome of this appeal will apply to all cantons. More recently, a Cour d’appel cantonale vaudoise has again stated that Uber chauffeurs are employees. Apparently, Uber did not appeal this second ruling (ATS, 2020; ATS/NXP, 2020; Swissinfo, 2020). Another country example that illustrates the importance of case law is Australia, where a “multi-factor test” is used that is set out in case law, but which does not always generate clear guidance, according to some observers (G. Golding, 2019, pp. 177-179).92 Nevertheless, some local authors seem to point towards the courts as a driving force for improvement (besides Golding, also see, for instance, R. Rideout, 2000, p. 421). Similarly, in the United Kingdom, the Trade Union Confederation has expressed opposition to the government’s plans to codify existing common law tests in legislation because this would “restrict the ability of the courts to adapt and apply the tests to new emerging forms of work, including platform work” (ILO Committee of Experts, 2020, p. 109)

Other authors, such as Anderson in New Zealand, on the other hand, argue in favour of more codification (G. Anderson, 2016, p. 116), as also done by the New Zealand Council of Trade Unions (NZCTU) (ILO Committee of Experts, 2020, p. 109). Indeed, the NZCTU has proposed to broaden the notion of an employee under the Employment Relations Act 2000 (ERA) to allow certain workers to be captured under this notion.93 More specifically, according to this proposal, section 6 of the ERA, which provides the criteria for the existence of an employment relationship, ‘could be extended to codify tests and factors in common law and add further (non-exhaustive) factors - “economic dependence” (to capture dependent contractors) and “imbalance of bargaining power”’ (New Zealand Council of Trade Unions, 2019, p. 21).

Which of these two approaches – case law or codification – is preferable remains an open question. It is clear, however, that the interaction between case law, legislation and administrative sources on the employment relationship may be particularly relevant. Some countries that have adopted legislation on the matter, such as Belgium, have mostly codified the principles from domestic courts’ decisions into statutory law.94 Similarly, in Germany, the employment contract has been statutorily defined in 2017 by transposing German case law on the matter (M. Schlachter, 2019, p. 230). Section §611a of the Civil Code stipulates that the employment contract obliges the employee in the service of another person to perform instruction-dependent work, that is externally determined and performed in a personally dependent capacity. The section, furthermore, details that a court must take all circumstances into account when determining an employment contract. If the actual execution of a contractual relationship shows that it is an employment relationship, the classification given by the parties to the contract is irrelevant.95

The same is, to some extent, true for California. The recent “AB5 law” builds on the California Supreme Court’s decision in Dynamex Operations West v. Superior Court of Los Angeles County.96 It establishes that the so-called “ABC” test should, in principle,97 be used by courts in the determination of the employment status for purposes of applying the Labor Code, the Unemployment Insurance Code and the wage orders of the Industrial Welfare Commission.98

Under California’s ABC test, businesses that seek to treat workers as independent contractors must show that: “(A) [t]he person is free from the control and direction of the hiring entity in connection with the performance of the work, both under the contract for the performance of the work and in fact. (B) The person performs work that is outside the usual course of the hiring entity’s business. (C) The person is customarily engaged in an independently established trade, occupation, or business of the same nature as that involved in the work performed.”99 It is not possible to rule out entirely that some platform operators might succeed in doing so. Apart from claiming a lack of control and direction, which we discuss below, certain platform operators might also succeed to meet criterion B, because they enable a wide range of activities. For instance, Taskrabbit enables platform workers to clean, mount furniture, deliver groceries and so forth. Therefore, its “usual course of business” may be more difficult to identify compared to other platforms. Moreover, for instance, if a plumber uses Taskrabbit to receive certain jobs, but in addition to the platform, also has “an independently established occupation” as a plumber, then it might be possible to maintain that the platform worker is independent under the ABC test. On the other hand, “if a worker has not independently decided to engage in an independently established business but instead was simply designated an independent contractor by the unilateral action of the [platform], then there is a high risk of misclassification and factor C may not be satisfied” (I. Lianos et al., 2019, p. 296). It remains to be seen whether, for instance, certain workers on a single platform end up being classified as employees, whereas other workers on the same platform are self-employed.

Arguably, this Act goes in the direction of “allowing a broad range of means for determining the existence of an employment relationship”, as indicated in Paragraph 11 of Recommendation No. 198, even if Recommendation No. 198 specifically mentions “national” rather than “State policies”. According to some scholars, the Act's adoption would entail that many platform workers would most likely be considered employees of the relevant platforms (B. Sachs, 2018). This is, in particular, the case for “non-entrepreneurial workers who are engaged in a platform's ordinary course of business” (K. Cunningham-Parmeter, 2019, p. 472). The adoption of the AB5 law has, furthermore, generated interest among labour-friendly politicians in States, such as New Jersey and New York (D. Rubinstein et al., 2020).

Platform operators, on the other hand, have tried to challenge the Act in numerous ways. The enterprises have argued that it violates the State and Federal Constitution and that, in any case, platform workers would not classify as employees even under this test (A. Marshall, 2020). In this respect, in February 2020, Uber and Postmates attempted, without success, to impose a preliminary injunction against the enforcement of the AB5 law, by suing the State of California (J. Schaedel et al., 2020).100 The most recent development in this case was related to a direct ballot initiative, Proposition 22. During the election of November 2020, Californian voters were asked to decide on adopting a ‘Protect App-Based Drivers and Services Act’. The Act was supported by many labour platforms. It aimed to ensure most platform workers not to be subject to the ABC test. These workers were to be classified as self-employed while the Act would attribute limited protections to them. The Californian voters approved the proposal (M. A. Cherry, 2021). As a result, the platform workers in question will be classified as self-employed, with certain specific protections applying to them, including non-discrimination safeguards and minimum earnings for effective working time. However, since the court rulings before the ballot indicated that these platform workers were indeed employees under the existing Californian ABC test, some platform operators might still be liable to a certain extent. This is because they might have failed to comply with the relevant labour regulations from the moment AB5 came into effect until when Proposition 22 superseded it.

The Californian AB5 law might also have an indirect impact in the Netherlands, where an independent commission seemed to have drawn inspiration from it to deal with employment classification (Commissie Regulering van Werk, 2020). It concerns a sensitive issue in the Netherlands, amongst others, because of their high percentage of solo self-employed workers.101 In the past, a ‘Labor Relationship Decision’, that had been issued in 2006, used to specify the difference between employment and self-employment status for tax and social security purposes.102 This decision has been replaced by the ‘Deregulation of the Assessment of Employment Relations Act’ of 2016.103 This Act arguably aimed at limiting the risk for principals, who relied on independent contractors, to be subsequently characterized as employers of disguised employees (C. Overduin, 2015). However, although this Act entered into force on May 1, 2016 (E. Swaving Dijkstra, 2016), its enforcement has been put on hold because of concerns about its clarity.104 Following a decision in 2018, the Dutch Government initiated a plan to replace the 2016 Act.105 A new law is expected to follow in 2021, as part of an integrated approach to labour market policy.106

As a part of this integrated approach, the government issued a report on the future of work. An independent commission reviewed, amongst others, the national policy on employment classification (Commissie Regulering van Werk, 2020, pp. 68-72). In this report, the commission recommends “modernising” the concept of subordination to ensure that the integration of a worker in the enterprise’s business is made of central importance. Also, whether the worker’s activities are part of the enterprise's regular activities should be emphasized as a relevant criterion (p. 70). According to the commission, these two criteria would not radically depart from the tax authority’s current guidelines on employment classification.107 Furthermore, the commission proposes a presumption of employment that would imply that workers providing labour personally in return for remuneration are employees, unless the employer proves that they are independent contractors.108 This, again, would be in the direction of implementing one of the measures described in Recommendation No. 198, namely the introduction of a legal presumption of employment “where one or more relevant indicators is present”. Recent documents indicate that the government might also be willing to establish a targeted presumption favouring platform workers’ employee status, if these workers fail to pass the test mentioned above (W. Koolmees, 2021).

Another example of general legislation that may have a paramount importance in the regulation of platform work is the Italian 2015 reform, as described above, and aimed at extending labour, employment and social protection to all workers whose work is organized by the other party, even if they do not meet all the requirements to classify as an employee under the definition of the Civil Code, as interpreted by courts.

Public administrators and labour inspectorates can also play an important role in enforcing the regulation in relation to platform work and the determination of employment status, in line with Paragraph 15 of Recommendation No. 198. Labour inspectorates started inquiries and issued administrative guidance concerning platform work in France, Denmark, and Sweden, among others (Z. Kilhoffer et al., 2019, pp. 117-118). The labour inspectorates have, furthermore, been very active in Spain, where numerous court decisions accepted the classification of platform workers as employees, as suggested by the inspectorates (A. Barrio, 2020). In Poland, the Solidarność trade union has asked the national labour inspectorate to audit Uber (T. Haipeter et al., 2020, p. 23), and in Belgium, a coordinated investigation has been launched into Deliveroo, which will be pursued in front of the courts.

These approaches make regulations aimed at directly involving labour inspectorates to fight employment misclassification all the more relevant. An example in this respect is Portugal’s Act No. 63/2013.109 This law established mechanisms to combat the misuse of service contracts. If labour inspectors find that the relationship between two parties bears an employment contract's features, they can draft a notice and request the employer to regularize the situation within ten days. If this does not occur, the relevant file is referred to the Public Prosecutor to bring an action to recognize the existence of an employment contract.

Similarly, to improve enforcement, Slovenian authorities took measures once they noticed an increased usage of civil law contracts in the period between 2013 and 2016. This increase was also due to the misclassification of employees as independent contractors (Eurofound, 2017). Therefore, lawmakers decided to amend the Labour Inspection Act (Zakon o inšpekciji dela) in September 2017. Under the amended Act, and of course subject to reconsideration by the labour courts, the Labour Inspectorate has the power to compel employers to enter into an employment relationship with a person illegally working under a contract for services.110

Apart from the examples above, certain other regulatory measures may also impact platform workers' employment status. This might be the case, for instance, of regulation on labour brokerage or private employment agencies. As discussed in Section 2, sometimes, the platform's user is regarded as the employer (e.g. Helpling in the Netherlands), with the platform being treated as a digitalised job-matching agency. Platform operators can, however, also provide their services by adopting the legal status of a temporary work agency, or another form of private employment agency (J.-Y. Frouin, 2020). Other regulation that can be relevant may concern telework and ICT-based mobile work, as well as casual and on-call work, as will be shown below.111

For example, the scope of the regulation about private employment agents in Queensland (Australia) is interesting. A person is deemed a private employment agent “if the person, in the course of carrying on business and for gain— (a) offers to find — (i) casual, part-time, temporary, permanent or contract work for a person; or (ii) a casual, part-time, temporary, permanent or contract worker for a person”.112 The reference to “contract work” and “contract worker” seems to suggest that also the brokerage of contracts for services through platforms may be subjected to regulation.

Some regulatory measures, which contain provisions on “casual”, “on-demand” or work with unpredictable work patterns, might also impact platform workers. The recently adopted EU Directive on Transparent and Predictable Working Conditions (TPWCD) provides an example in this regard.113 This legal instrument is a step forward in ensuring some minimum predictability to workers with unpredictable work schedules (A. Aloisi et al., 2019). This is done, among other things, by mandating a reasonable notice period for assignation and cancellation of shifts, as well as predetermined reference hours and days, and by introducing the right to request a transition to more stable employment. According to recital 8 of this Directive, platform workers could benefit from the abovementioned minimum rights, if they fall within the notion of ‘worker’ under article 1 of the Directive. This article requires the existence of “an employment contract or employment relationship, as defined by the laws, collective agreements or practice in force in each Member State with consideration to the case-law of the Court of Justice.” As a result of the choice to use the employee-concept under national law for the Directive’s scope of application, albeit “with consideration to the case-law of the Court of Justice”, a vast part of platform workers risk being excluded from the protection accorded by the Directive.114 Commentators have indeed observed that, considering that the Directive does not tackle the issue of misclassification, “most self-employed persons engaged in gig work will remain outside its scope of protection”, unless Member States broaden their understanding of the employment relationship under national law (B. Bednarowicz, 2019, p. 607). As mentioned in Section 4 below, several national courts in the EU have been ready to recognize the existence of an employment relationship between some platforms and their workers. This, coupled with a reference to the EU Court of Justice's jurisprudence, which ordinarily adopts a broader understanding of “employment” than national legislation, makes the Directive extremely relevant for platform work.115 A recent order by the EU Court of Justice, in which the Court reiterated its existing case law and applied it to a parcel delivery undertaking, does not alter this conclusion. The Court did not rule in favour of the reclassification of the claimant as a worker under the Working Time Directive, considering that the worker at hand enjoyed a very high degree of organizational and work autonomy.116

Furthermore, the potential relevance of legislation on casual work for platform workers has been confirmed in the 2020 General Survey of the CEACR. According to this survey, platform work is considered, on many occasions, a form of casual work or zero-hours contract arrangement (ILO Committee of Experts, 2020, p. 140). However, in order for national legislations on casual work to be applicable to platform work, an employment relationship needs to be established first. Whenever the classification as an employee has been reached, platform workers may benefit from the provisions that put minimum predictability in place for casual workers. Some legislative examples, which provide limitations to the use of zero-hours contracts, may grant some minimum predictability to platform workers, if they classify as an employee.117

In Brazil, Act No. 13467/2017 introduced a new form of work, the “intermittent [employment] contract”.118 This type of contract allows employers to engage workers to provide their services on a non-continuous basis, as an employee. The reform would help employers retain a flexible workforce ready to perform services when necessary. Workers’ salary cannot be lower than the minimum hourly wage of other company employees who work in the same position under an intermittent or regular employment contract. They are also entitled to a full range of benefits, such as severance payments, social security payments and bonuses. Together with workers’ access to these benefits, these reforms may enable employers to have more contractual flexibility; this could arguably lead them to hire more intermittent workers instead of offering full-time employment (J. Martinusso, 2018). It remains to be seen whether platforms will make use of these intermittent contracts and, as such, they will refrain from making recourse to a (bogus) self-employed workforce.

Regulation on telework, distant work or remote work may also be important, as mentioned. A new chapter was added to the Russian Labour Code in 2013 on distinctive features of employees' remote work.119 Through this new chapter, regulation is provided to “on-distance”, “remote” employment relationships “performed outside of employer’s premises and outside a standard workplace controllable by an employer,” which are mediated by digital technologies, especially by the internet. The workers covered by these new provisions are employees who receive a salary and right to paid leave; while, at the same time, in principle, being allowed to determine their own work hours and rest time. According to some researchers, “the vast majority of online platform workers, and workers of the mobile apps, continue working outside of the scope of this legal provision, despite the fact that their relationships with online platforms, or with online clients, often have all features of the remote employment relationship” (M. Aleksynska et al., 2020). According to the same researchers, this can relate to the fact that this provision does not explicitly regulate platform work and targets traditional employment relationships instead, whose obligations are performed through IT. This is certainly something to reflect upon.

The Brazilian Act No. 12,551 of 15 December 2011 has added a very relevant provision in the country’s Consolidation of Labor Laws.120 It explicitly states that the telematic and computerized means of command, control and supervision are, for legal subordination purposes, equal to the personal and direct means of command, control and supervision of others’ work. This amendment was initially passed to ensure that teleworkers be considered employees,121 but it may also benefit platform workers who wish to establish an employment relationship.122

4.1. Principle of primacy of facts

The principle of primacy of facts is arguably one of the most important contributions of Recommendation No. 198 to the already pre-existing body of international labour standards. This principle is broadly adopted and widely applied by national courts (see ILO, 2016, p. 262). Our case law analysis has not found any rulings in which courts cling to the contractual terms of the agreement to the detriment of an assessment based on the actual circumstances of the case. Moreover, some courts explicitly refer to Recommendation No. 198 in applying the primacy of facts principle. For instance, in Uruguay, the Juzgado Letrado del Trabajo referred to this instrument to sustain its finding of the existence of an employment relationship between Uber Technologies and the worker(s).126 In its appeal, Uber explicitly disputed the extent to which the first tribunal relied on ILO Recommendation No. 198. However, the Uruguayan Tribunal de Apelaciones de Trabajo rebutted Uber’s argument, holding that “[t]he Chamber agrees with the first instance ruling that, at the present time and at the time of the occurrence of the facts in question, ILO Recommendation No. 198 should be considered as the theoretical framework applicable in Uruguay when there is a dispute over the qualification of a legal relationship involving employment.”127 Other courts derive the principle of primacy of facts from, amongst others, the EU Directive on transparent and predictable working time,128 or national statutory law.129 Importantly, this principle is not only applied when plaintiffs claim to be full-fledged employees. Litigation in which plaintiffs claim to be dependent contractors, such as in Ontario, or “limb (b) workers”, such as in the United Kingdom, is likewise underpinned by this principle.130

4.2. Flexibility of work schedules

Paragraph 13 of Recommendation No. 198 mentions the facts that work is “carried out within specific working hours or at a workplace specified or agreed by the party requesting the work” or requires “the worker's availability”, as some of the possible indicators of the existence of an employment relationship. The possibility of platform workers to set their schedules and not be formally obliged to log on the platforms at a fixed time of the day or the week, something that could be referred to as ‘flexibility of work schedules’, is arguably related to those indicators. So far, this flexibility has mostly been used by national courts as a primary ground to classify platform workers as independent contractors. This happened, for instance, in Brazil where both the Superior Tribunal de Justiça and a Tribunal Superior do Trabalho ruled that the Uber drivers’ flexible work schedule does not allow for them being reclassified as employees.131 Also, the labour tribunal in Turin, Italy, rejected some food-delivery riders' reclassification claim by extensively relying on the flexibility that they enjoyed in terms of work schedules.132 A similar argument, centred on the lack of fixed working hours, was upheld by the labour tribunal in Milan (Z. Kilhoffer et al., 2019, p. 115).133 In some cases, even if the flexibility of work schedules does not constitute the main reason underlying the rejection of the existence of an employment relationship, it has still been used as an indicator in this direction, such as in the case of Mr Michail Kaseris v. Rasier Pacific V.O.F., in Australia.134 The same is true for some decisions in the United States, including interpretations issued by administrative bodies.135

However, this flexibility does not constitute an insurmountable obstacle in other jurisdictions. A good illustration can be found in the French case law. Whereas the cour d’appel de Paris had stated in 2017, in relation to Deliveroo, that the ability to work whenever one wants stands in the way of an employment relationship,136 the same court noted in January 2019, in relation to Uber, that the freedom to connect and, consequently, the free choice of working hours does not in itself exclude an employment relationship, as long as it is demonstrated that when the drivers connect to the platform, they are integrated into the service, as organised by the company, which gives them instructions, controls their execution and exercises disciplinary power over the worker.137 In March 2020, the Cour de cassation confirmed the Parisian Court of Appeal’s decision of January 2019, according to which Uber drivers are employees.138 Choosing one’s work schedule by logging on and off the app does not exclude the existence of an employment relationship, if the driver becomes an integrated part of the platform’s services once logged on.139 Similarly, according to a recent Belgian decision, it is not the ability to refuse tasks that determines working time flexibility but the ability to determine one’s own working time once the work has been accepted. In a 2020 landmark case, the Spanish Tribunal Supremo also put platform workers’ alleged flexible scheduling under close scrutiny. First of all, the freedom to schedule work would not in any way preclude the existence of a contract of employment. Moreover, for Glovo couriers, the theoretical freedom to choose time slots was deemed to be limited. According to the Supremo, the platform uses a “scoring system” that considers the customer’s ratings, the efficiency demonstrated when executing gigs, and the performance of services during peak hours. When couriers cannot work during peak hours, their scores drop, making it harder to access the time slots with the highest customer demand. Therefore, the theoretical freedom to schedule time slots is quite different from the actual freedom, in the court’s view.140 Beyond the Spanish Tribunal Supremo’s decision, mentioned above, the argument that flexibility in work schedules prevents the establishment of an employment relationship has likewise been refuted by some other courts in Spain, in particular, concerning the platform Glovo. The assertion that its couriers can freely schedule their work and can freely reject orders has to be contextualized, according to the Tribunal Superior de Justicia de Madrid. The platform still decides when to allow workers to work on the app based on the anticipated demand and according to the algorithm’s metrics. The freedom to choose a time slot is significantly tempered.141 For instance, even if one is not directly penalized for rejecting an order, one’s rating of “excellence” decreases when this is done, thus preventing access to the best and most advantageous time slots and limiting this freedom. Based on similar reasoning, the Tribunal Superior de Justícia de Catalunya has overturned a prior decision by the Juzgado de lo Social de Barcelona. Glovo couriers are employees, according to the higher court, even if the workers have no fixed work schedule, predetermined rules to take holidays and so forth. These characteristics are a normal effect of an employment relationship rather than foundational elements that constitute one. Glovo, according to the court, exercises control over riders’ schedules through variable remuneration and methods of rating and evaluation. If workers want a higher income, they must work in high-value time slots, which only become available on the basis of these evaluations.142

The Netherlands offers another example of how working time flexibility is not incompatible with employment status. A court in Amsterdam had ruled in July 2018 that Deliveroo couriers are self-employed, in part because of their flexibility with regard to working time. The court argued that the high performing couriers, indeed, gain “priority access” to time slots. However, this did not entail that the less performant couriers would lose their ability to make reservations – they were only able to do so after the high performing couriers. Although this dynamic could indicate “(a form of) authority”, this was not enough to establish subordination under the Dutch Civil Code, according to this judgement.143 Another court of first instance in Amsterdam contradicted the first court’s ruling in July 2019. It argued that Deliveroo, in fact, did limit couriers’ freedom of working time.144 The couriers can, therefore, still be considered bound by an employment contract, despite the significant amount of freedom enjoyed.145 The Court of Appeal followed this second court’s reasoning in a recent ruling from February 2021. Based on all the facts, the Court argued, the only circumstance that indicates the couriers’ self-employed status is their freedom to organise the work. However, in the opinion of the Court, this flexibility was not incompatible with an employment relationship between Deliveroo and the couriers. Deliveroo’s decision to transition from a “self-service booking tool”, in which couriers had to preregister for time slots, to a “Free Login”-system, allowing workers to come online anytime, did not alter the Court’s conclusion. Considering all the elements of the relationship between the couriers and the platform, the former were to be regarded as employees, according to this judgment.146

The extent to which the freedom with which platform workers organise their work(ing) (time) can or cannot be perceived as a matter of secondary importance is highly dependent on the national legal system’s conceptualisation of the employment relationship. Some legal systems, for instance, retain the idea that a so-called ‘mutuality of obligation’ is essential for an employment contract. This implies that “there must be mutual obligations on the employer to provide work for the employee and on the employee to perform work for the employer”147 (A. Adams and J. Prassl, 2018, p. 27). Compared to other common law jurisdictions, a particularly strict reliance on mutuality of obligation can be observed in the United Kingdom (N. Countouris, 2015; V. De Stefano, 2016b).148 Nevertheless, this is also an important element in some civil law systems, as demonstrated by the Italian tribunal judgments cited at the beginning of this section. Also, in Germany, a Landesarbeitsgericht in Munich emphasized that a framework agreement, i.e. the platform’s terms and conditions, which does not create an obligation to perform work, is not an employment contract. The workers were asked to visit retailers to verify how the products were displayed, taking pictures and filling in questionnaires.

The fact that the platform worker performed a high number of orders, and was even incentivised to accept orders due to the platform's functioning did not lead to another conclusion.149 In this case, the court made extensive reference to the fact that the worker “was [still] free to decide whether and, if so, which orders he was processing” for the platform business in question.150 It held that this degree of sovereignty over time was “absolutely unusual for an employment relationship”.151 This was, moreover, explicitly indicated within the framework agreement ("The contractor is free to accept an available order at any time, there is no obligation to do so") and, according to the court, the actual circumstances did not depart from this provision. This decision by the Landesarbeitsgericht was overturned by the Federal Labour Court, the Bundesarbeitsgericht. The Court recognized an employment relationship in the case at hand. As to the worker’s flexible working time, it decided that although the worker was free to accept or decline work, the platform operator induced the worker to carry out the work, using an incentive system. In combination with other elements, such as the detailed instructions, this led the Court to rule in favour of the worker.152 A similar platform in France, called Clic and Walk, was also found to be an employer.153

Arguably, the centrality of this flexibility of work schedules in all these judgments shows how some of the traditional indicators of the existence of an employment relationship could be ill-suited to face the emergence of highly casualized forms of work such as platform work; concentrating only on this element may lead to the decision of excluding from protection exactly the most precarious and unstable workforce (A. Adams et al., 2015).154

As such, in the spirit of constantly updating the existing national policies concerning the employment relationship, advocated by the Employment Relationship Recommendation, 2006 (No. 198), it could be possible to re-examine the role that flexibility of work schedules should play in the determination of the existence of the employment status in modern labour markets.

4.3. Control through technology

The existence of “control” of the employer over the work is one of the, if not “the” most prominent element necessary to establish an employment relationship in many legal systems around the world (in civil law jurisdictions, this ordinarily corresponds to the technical “subordination” of employees to employers) (see ILO, 2016, p. 11). The possibility of the platforms exerting control over the workers, particularly through technological tools such as algorithms, rating systems and geo-localization devices, is a crucial element in many judicial and administrative decisions – worldwide – on the employment status of platform workers. The Republic of Korea offers a good example. The Ministry of Employment & Labour classified food delivery couriers working for “Yogiyo” as employees. The workers had argued that Yogiyo “exercised substantial supervision and control over them as workers” (K. H. Ryoo et al., 2019). The National Labour Relations Commission of Korea has also ruled that a former driver for a transportation platform called “Tada” was an employee; its “decision centered on the degree of control and direction imposed on the driver” (B. Gu and H. C. E. Yu, 2020).

It is interesting to notice how also court judgments not directly related to the application of employment regulation could offer useful insights to a labour law analysis in this respect. This is arguably the case of the judgment Asociación Profesional Elite Taxi v. Uber Systems Spain SL, delivered in December 2017 by the Court of Justice of the EU. In this judgment, concerning the applicability of national transportation regulations to Uber, the Court ruled that Uber services are, in fact, transport services, instead of mere “digital” information society services; this allows EU Member States to regulate these services at the national level. In taking this decision, the Court analysed the amount of control that Uber exercises on its drivers. According to the Court of Justice, this control is exercised when selecting the drivers and the vehicles they use and when the service is performed.155 The Court found that Uber does not only fix the maximum fare by using its eponymous app, but also provides financial rewards in case of accumulation of trips. The platform uses the rating function to give access to better-paid jobs and even exclude poorly-rated drivers from the platform. As already mentioned, this judgment does not directly concern the issue of Uber drivers’ employment status. It is nonetheless important to notice that, in an Opinion for this case, the Court’s Advocate General had observed: “[i]ndirect control such as that exercised by Uber, based on financial incentives and decentralised passenger-led ratings, with a scale effect, makes it possible to manage in a way that is just as — if not more — effective than management based on formal orders given by an employer to his employees and direct control over the carrying out of such orders.”156

The Court’s ruling and the Advocate General’s Opinion have arguably informed certain national courts’ views on the matter. A Spanish Tribunal Superior de Justicia, for example, used both sources in its arguments concerning the classification of Glovo couriers.157 This is not to say, of course, that all European national courts interpret the Court of Justice’s finding that Uber “exercises decisive influence” over its drivers as a proxy for subordination, or even take it into account. There could still be many reasons why Uber would not be considered an employer pursuant to national law. Taxi companies in Brussels, for instance, sued Uber before a Belgian tribunal de l’entreprise to claim that the platform should be banned for breaches of transportation law. The alleged misclassification of Uber drivers as self-employed workers was only filed as an additional claim, which is why the case was not brought before a labour court. This, however, did not dissuade the tribunal from ruling that Uber drivers were self-employed. To that end, the tribunal explicitly differentiated between Uber drivers and Deliveroo couriers, who had been considered employees by an earlier decision of an administrative body.158 In making the comparison, the tribunal mentioned, inter alia, that Uber drivers have a transport licence, contrary to Deliveroo couriers. Furthermore, whereas, in the case of Deliveroo, geo-localisation supposedly only serves for the client to check up on the arrival of the food, the geo-localisation of Uber drivers is indispensable to the functioning of Uber. Quite surprisingly, the tribunal went on to argue that not only the Uber driver is tracked, but also the passenger and “it cannot be seriously argued that Uber BV would exercise hierarchical control over all the [passengers]; although, they are just as geolocatable as the [drivers], provided that both are connected to the platform”.159 In this respect, the tribunal seems to argue that since passengers are not hierarchically controlled, therefore, also drivers cannot be considered hierarchically controlled.160 This finding goes in the same direction as a court decision issued in Florida (United States), according to which Uber does not provide any “direct supervision” but only leaves it up to the passengers to evaluate or supervise the drivers through a rate review system.161 This is an observation that a Court of Appeal in France also made.162 Similarly, the Employment Court in Auckland, New Zealand, albeit accepting that Uber is in the passenger transport business, ruled that a driver was, in fact, an independent contractor.163 Among other things, the Court ruled: “work was not directed or controlled by UBER beyond some matters that might be expected given [the driver] was operating using the Uber ‘brand’”.164 It should be mentioned, however, that the Court pointed out that this decision was based on “an intensely fact-specific”165 interpretation.

In South Africa, instead, in a case brought against Uber for unfair dismissal, the Commission for Conciliation, Mediation and Arbitration (CCMA) found the existence of an employment relationship. The CCMA Commissioner pointed out that: “even though there is no direct or physical supervision, control is exercised through technology, to the point that even the movement of the cell phone can be detected, indicating reckless driving”.166 The French Cour de cassation also drew heavily on the element of geo-localisation to overturn a decision by the cour d’appel about the Take Eat Easy platform. According to the Cour de cassation, a company that uses a platform and application to connect users, thereby tracking their location in real-time, to the extent that the total number of kilometres travelled becomes evident, and wielding power to sanction the worker cannot engage those workers as independent contractors.167 In a more recent case, the French Cour de cassation further acknowledged Uber's control via its algorithms by setting the fares, overseeing the acceptance of rides and imposing a certain route to be followed by the driver.168 The power to sanction drivers was also exercised, for instance, through fare adjustments in case drivers choose an “inefficient route”, through Uber’s ability to deactivate the drivers, and through the order-cancellation rates applied to them. Likewise, the Labour Appeal Tribunal of Montevideo (Uruguay) ruled that Uber directs and controls all of a driver’s performance, and its supervising activities represent a “form of exercise of the sanctioning power typical of an employer”.169 Analogously, an administrative body in Belgium decided that Uber’s ability to give instructions, monitor their compliance, and exclude a driver from access to the application in the event of non-compliance with these instructions reveals a hierarchical control incompatible with the legal classification of an independent work relationship.170

Additionally, the Tribunal Superior de Justicia de Madrid noted that the geo-localisation system of Glovo exercises effective and continuous control over the courier's activity.171 The same is true for the Spanish Tribunal Supremo, which mentions that the couriers receive instructions and are subject to a “permanent control system”. The court explains: “[t]he company has established a scoring system which, among other factors, is fed by the final customer's assessment. The establishment of production activity control systems based on customer assessment is a favourable indication of the existence of an employment contract.”172 A regional labour court in Brazil made a similar argument. The court also argued that technology could enable platforms to enhance their overall control over workers’ activities. According to the judgment, control by everyone and no one can be more effective than control by a dedicated manager.173

In a recent case in Canada, the Ontario Labour Relations Board also seems to point out how technological developments can constitute an additional layer of control over the work performed. Much attention was given to the fact that the platform, Foodora, can control its couriers' work. The board observed: “in addition to tracking and reporting issues with couriers, and investigating issues, the use of Global Positioning System (“GPS”) technology is an additional layer of control.”174 It also highlighted that “the advancement of technology – algorithms, GPS, automated alerts, SMS communications – allows Foodora to control the operation with minimal human interaction. This does not mean Foodora does not closely supervise the couriers.”175 Another case in Pennsylvania (United States) illustrated how courts can have a very different understanding of the extent of platforms’ control. A district court had concluded in 2018 that the control test in the case at hand “weighs heavily in favor of “independent contractor” status” and decided to uphold UberBlack drivers’ classification as self-employed workers.176 The United States Court of Appeals for the Third Circuit decided to remand the matter for further proceedings. The appellate judges were not entirely convinced that UberBlack drivers are not subject to control under the Fair Labour Standards Act.177 They highlighted, in fact, that “Uber deactivates drivers who fall short of the 4.7-star UberBLACK driver rating and limits the number of consecutive hours that a driver may work” (K. Dailey and E. Mulvaney, 2020). Even in the same legal system, therefore, courts can construe the element of control exercised through technological tools very differently. In a recent case related to the attribution of a driver’s unemployment insurance benefits, the Supreme Court of New York State ruled that there is substantial evidence indicating that Uber can exercise sufficient control over their drivers to establish the existence of an employment relationship.178 The Court, among other things, argued that Uber ‘provides a navigation system, tracks the drivers' location on the app throughout the trip and reserves the right to adjust the fare if the drivers take an inefficient route. Uber also controls the vehicle used, precludes certain driver behavior and uses its rating system to encourage and promote drivers to conduct themselves in a way that maintains "a positive environment" and "a fun atmosphere in the car."’179 Similarly, in the United Kingdom, the Supreme Court granted some Uber drivers the status of “limb (b) workers” by observing, among other arguments: “the ratings are used by Uber purely as an internal tool for managing performance and as a basis for making termination decisions where customer feedback shows that drivers are not meeting the performance levels set by Uber. This is a classic form of subordination that is characteristic of employment relationships.”180 Tech-enabled control will most likely be more and more relevant in the future, as forms of management-by-algorithm and algorithmic control spread much beyond platform work (V. De Stefano, 2020).181

4.4. Equipment and other inputs

One of the indicators that Paragraph 13 of Recommendation No. 198 identifies as potentially pointing towards employment status is the fact that the work “involves the provision of tools, materials and machinery by the party requesting the work”. Whether or not the worker wears a logo or uniform of the principal, in fact, is an indicator that is taken into account in certain jurisdictions. This was particularly relevant in many judgments on platform work issued by courts around the world. A Chilean judge, for instance, ruled that Uber did not employ its drivers, amongst other factors, because they do not wear a uniform, something that the claimant had apparently argued.182 Other judges also reference the fact that platform workers do not wear uniforms.183 The same stance was also taken in a case in Australia, where the Fair Work Commission decided that the fact that the worker in question was not permitted to display any of the company’s or its affiliates names, logos or colours on his vehicle was a factor that weighed against the existence of an employment relationship.184 The same ruling also considered the fact that the worker was required to provide his own capital equipment (i.e. his own vehicle, smartphone and wireless data plan) to use the Partner App and charge fares was another indication of the absence of an employment relationship.185

Similar findings have been made in the United States. One of the six so-called “Donovan factors”186 to determine the existence of an employment relationship under the USA’s Fair Labor Standards Act is “the alleged employee's investment in equipment or materials required for his task, or his employment of helpers.”187 According to the district court in Pennsylvania, this factor strongly favours independent contractor status since Uber drivers buy their own car, thereby making “significant capital investments”. An opinion letter of the U.S. Department of Labor seems to share this view. Furthermore, it adds that the platform operator's investments in its own platform do not make a significant difference with regard to this indicator “because they are not investments in the work the service providers perform. [...] To be sure, the service providers rely on [the platform’s] software to quickly obtain jobs, but that reliance only marginally decreases their relative independence, because they can use similar software on competitor platforms.”188

Not all courts, however, share this point of view. In a case about Foodora, for instance, the Australian Fair Work Commission found that “[t]he applicant did not have a substantial investment in the capital equipment that he used to perform his delivery work.”189 Other courts compare the workers' investment in their tools to the significance of the technology investment of companies in the provision of the service. For example, a decision in Uruguay mentions that by providing the application to its drivers, Uber provides them with the tools to work.190 Similarly, the Ontario Labour Relations Board has argued that “[w]hile the tools used to make deliveries are supplied by both the courier and Foodora, the importance of the App cannot be ignored. It is the single most important part of the delivery process and is a tool owned and controlled by Foodora.”191 A court in Madrid, in weighing the relevance of the tools provided by the parties, also observed: “it is sufficient to compare the advanced technological instruments of which the defendant is the owner (digital platform and computer applications), and the very insignificant means provided by the applicant (mobile phone and motorcycle).”192 Also according to the Spanish Tribunal Supremo, the essential tool of production in this activity is the digital platform of Glovo, not the motorbike or mobile phone of the worker.193 The same is true for Uber, according to the Supreme Court of the United Kingdom.194 Although many courts pay attention to the ownership of tools,195 others seem to consider it as relatively insignificant, for example, because a bike is an object commonly used in many people’s daily life, rather than an investment.196

4.5. Substitution clauses

Recommendation No. 198 mentions as one of the possible indicators of an employment relationship's existence, the fact that work “must be carried out personally by the worker”. Traditionally, employment contracts have consistently been seen as arrangements where the person of one of the contracting party, the worker, is an essential feature (so-called “intuitu personae contracts”). This issue is particularly significant when examining case law about platform work. Some platforms, in fact, grant their workers the possibility to be substituted by other workers. The terms and conditions between the worker and platform often contain so-called substitution clauses, that detail under which conditions the workers can exercise this right. The substitute should, for instance, be covered under the primary worker’s third-party liability insurance. These clauses have become a battleground in some litigation, in particular in the UK. Platform operators argue that these clauses indicate workers’ status as independent contractors, while workers argue that such clauses are only marginally used, if at all, and should not preclude their access to social protection. The matter of personal performance of the work is not only relevant to the UK’s legal system. For example, a Brazilian judge decided that some Uber drivers were not employees, as they did not perform the work personally, considering how they can have others do the driving for them.197

Most of the decisions on this topic are, nevertheless, from the UK. The Central Arbitration Committee noted that Deliveroo couriers do, in fact, have a genuine “substitution right”. As a consequence, the couriers do not personally undertake to work for Deliveroo, which entails that the couriers are not legally considered as ‘workers’ under the Trade Union and Labour Relations (Consolidation) Act 1992 (TULR(C)A) or the Employment Rights Act 1996.198 As such, they are excluded from most collective labour rights, minimum wage, working time and anti-discrimination protection attached to the so-called “limb (b) worker status”, i.e. one of those hybrid statuses between employment and self-employment mentioned in Section 1 above. The High Court of Justice has confirmed the Central Arbitration Committee’s decision.199 It should be stressed, however, that these clauses do not always wholly neutralize workers’ claims to a “limb (b) worker status”. An employment tribunal ruled in favour of some workers, despite such clauses. It was found that the workers in question did not have the right to substitute in practice.200 In another case, an employment tribunal pointed out that the substitution clauses at hand did not confer “an unfettered right to substitute”.201 The applicants, thus, legally classified as limb (b) workers. 

This balancing between the theoretical right to substitute and its functioning in practice can also be found in other countries, such as Spain. The Juzgado de lo Social in Madrid, for example, referred to a ruling by the Supreme Court from 1986, in a case in which a company had allegedly provided workers with the possibility to substitute oneself on the basis of a substitution clause.202 According to the Supreme Court, this had seemed intended to distort the true nature of the employment contract, rather than actually granting such a possibility to the workers. The lower court seemed to argue that the same was also true concerning Deliveroo. A Juzgado de lo Social in Valencia came to a similar conclusion. By referring to earlier case law, the court pointed out that the substitution clause in question seemed to lack “effective relevance in the execution of the contract”.203

The relevance of substitution clauses being often closely related to the contract's wording is another matter that needs to be assessed in light of the “primacy of facts” principle. Once again, thus, this principle, enshrined in the Employment Relationship Recommendation, 2006 (No. 198), proves to be an essential and crosscutting feature of the judgements around employment status.